CodeGoat24/UnifiedReward-7b
Updated
•
98
•
1
video_id
stringlengths 8
11
| video
stringlengths 22
25
| conversations
listlengths 2
2
|
|---|---|---|
seed42_328 | videos/HRA_part0/328.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a temperate deciduous forest, a professional lawyer, dressed in a gray suit and carrying a leather briefcase, wipes his brow as he pauses from his hike. The lush green forest, filled with tall trees and sunlight filtering through the leaves, provides a serene backdrop. The lawyer, wearing a gray blazer, button-down shirt, and khaki pants, is shown wiping his face with a handkerchief, revealing a moment of respite during his outdoor excursion. The peaceful atmosphere of the forest is highlighted by the dappled sunlight and the vibrant greenery surrounding the well-dressed man.]"
},
{
"from": "gpt",
"value": "Good. No obvious issues."
}
] |
seed42_2019 | videos/HRA_part0/2019.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A farmer, dressed in a straw hat, blue shirt, and khaki shorts, playfully swims in a clear, freshwater pond, holding a rabbit securely above the waterline. The farmer, now wearing a wide-brimmed hat, blue button-down shirt, and khaki shorts, gently cradles the rabbit, which sits comfortably in his hands, as they both float on the surface of the water. The idyllic scene is set against a backdrop of lush greenery and clear blue skies, capturing the essence of a peaceful rural life.]"
},
{
"from": "gpt",
"value": "Normal. The character's movements in the video are slightly unsmooth, and the rabbit's movements are also slightly unsmooth."
}
] |
seed46_1686 | videos/HRA_part2/1686.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A fluffy, baby panda with distinctive black and white fur crawls across a lush, green forest floor, surrounded by vibrant vegetation and bamboo shoots. It moves with a curious, almost playful demeanor, its tiny paws gripping the mossy ground. The scene is filled with natural light filtering through the dense canopy above, casting dappled shadows on the panda as it explores its environment.]"
},
{
"from": "gpt",
"value": "Normal. The limbs of the panda in the video are not fully generated, there are missing parts, the ears are problematic, and the paws and tail are not generated."
}
] |
seed42_2024 | videos/HRA_part0/2024.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A young, bespectacled female student sits cross-legged on a clean, white surface, diligently hemming a pair of black pants with a focused expression. She wears a grey cardigan over a white collared shirt and blue jeans, her attire casual yet neat. The environment is bright and minimalistic, with ample natural light illuminating her task. Tools of her activity, such as scissors, thread, and a needle, are neatly arranged within reach, highlighting her attention to detail and organization. The scene captures her absorbed in the sewing process, her hands skillfully manipulating the fabric, demonstrating a blend of practical skill and academic studiousness.]"
},
{
"from": "gpt",
"value": "Good. No obvious issues."
}
] |
seed42_1675 | videos/HRA_part0/1675.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A small, spiky European hedgehog leaps into the air, defying expectations as it jumps over a miniature hurdle, showcasing its unexpected agility. The hedgehog, with its typical brown and white quills, lands softly on a lush green surface, demonstrating its natural ability to navigate obstacles. The scene is set against a simple, uncluttered outdoor backdrop, allowing the hedgehog's unique action to be the focal point of the video.]"
},
{
"from": "gpt",
"value": "Bad. The hedgehog in the video walks unsmoothly and its movements are uncoordinated."
}
] |
seed46_1969 | videos/HRA_part2/1969.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A yellow Labrador dog, wearing a red collar and tags, happily swims in a natural hot spring pool, surrounded by lush green vegetation and a wooden dock. The water, clear and inviting, gently laps at the dog's fur as it paddles with ease. The dog's joyful expression and the serene environment of the hot springs create a peaceful and picturesque scene, enhanced by the vibrant greenery and the warm, bubbling water.]"
},
{
"from": "gpt",
"value": "Normal. The dog walks basically normally in the video, but there is slight uncoordination in its movements."
}
] |
seed44_2921 | videos/HRA_part1/2921.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a surprising yet harmonious scene, a veterinarian, dressed in a khaki safari outfit, carefully tends to a young rhinoceros in a spacious, fenced enclosure, with lush greenery visible in the background. The compassionate doctor gently examines the rhino, which is lying on a sandy surface, seemingly at ease. The rhinoceros, displaying signs of improvement, attempts to crawl forward on its front legs, indicating a recovery from a recent ailment or procedure, all under the watchful eye of the dedicated wildlife doctor.]"
},
{
"from": "gpt",
"value": "Bad. The person in the video has a twisted body, distorted face, and twisted hands; the rhinoceros has twisted legs, a twisted head, and a twisted body."
}
] |
seed44_2285 | videos/HRA_part1/2285.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A lawyer, dressed in a spacesuit adorned with patches and logos, embarks on a spacewalk to repair a satellite. She glides through the cosmos, the Earth visible in the distance, her movements methodical and confident. Her gloved hands, detailed with NASA symbols, work diligently on the intricate mechanisms, a tool kit secured to her waist. The vastness of space surrounds her, stars twinkling in the dark expanse, as she focuses on her task with the precision of her legal expertise. Throughout the sequence, her spacesuit helmet reflects the celestial backdrop, encapsulating the unique juxtaposition of her profession and this extraordinary setting.]"
},
{
"from": "gpt",
"value": "Good. Basically correct."
}
] |
seed42_1604 | videos/HRA_part0/1604.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A young student, dressed in a gray T-shirt and dark green shorts, carefully loads dishes into a stainless steel dishwasher in a modern kitchen. The room is brightly lit, casting clear shadows and revealing the intricate details of the kitchen, including the tile backsplash and granite countertops. The student, with short brown hair and glasses, methodically organizes the plates, bowls, and silverware, demonstrating responsibility and attention to detail. Throughout the process, the student maintains a focused demeanor, occasionally pausing to ensure each item is securely placed, all within the well-lit, contemporary kitchen environment.]"
},
{
"from": "gpt",
"value": "Normal. The actions of the person organizing the tableware in the video are not continuous and there are abrupt changes."
}
] |
seed44_1641 | videos/HRA_part1/1641.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A rugged, bearded man wearing a gray shirt, khaki cargo pants, and a brown belt pauses on a rocky trail, surrounded by lush greenery and towering trees. He leans on a wooden hiking stick, looking intently at the dense forest before him. Moments later, he hikes uphill, carrying a backpack and water bottle, with the sunlight filtering through the leaves overhead. Then, he stops to observe his surroundings, his weathered hands gripping the hiking stick, and the rich green foliage providing a stunning backdrop. Finally, he hikes through the dense wilderness, his attire and equipment suggesting a seasoned outdoorsman, and the vibrant greenery of the forest enveloping him in its beauty.]"
},
{
"from": "gpt",
"value": "Bad. The character in the video walks unsmoothly and uncoordinated on the rocky path."
}
] |
seed44_1819 | videos/HRA_part1/1819.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A middle-aged teacher, wearing a gray jacket over a button-down shirt, shakes vigorously, appearing disoriented, with eyes closed and mouth agape, against a classroom backdrop featuring a whiteboard filled with algebraic equations. Moments later, she is seen in profile, shaking her head from side to side with eyes closed, suggesting intense emotion. Then, she shakes her head again, her expression reflecting determination, as she stands against a classroom wall adorned with a periodic table poster. Finally, she looks directly at the camera, shaking her head in disbelief, her demeanor and attire consistent throughout the scenes.]"
},
{
"from": "gpt",
"value": "Bad. The woman's movements in the video are very unsmooth and uncoordinated."
}
] |
seed44_2513 | videos/HRA_part1/2513.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a surreal juxtaposition, a distinguished lawyer, clad in a classic black suit and white shirt, gently caresses the metallic surface of a Mars rover model placed on a wooden table. The room is dimly lit, casting dramatic shadows across the scene. The lawyer's face reflects a mix of fascination and professional poise, while the rover, adorned with intricate details and wheels designed for extraterrestrial terrain, stands as a symbol of human innovation and exploration. This intimate interaction between the legal world and space technology creates a thought-provoking narrative about the convergence of seemingly unrelated fields.]"
},
{
"from": "gpt",
"value": "Good. Basically no issues."
}
] |
seed46_1267 | videos/HRA_part2/1267.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A beaver, clad in a brown fedora and matching suspenders over a striped shirt, tap dances on a wooden stage under a spotlight. Behind him, a red velvet curtain provides a dramatic backdrop. He executes a series of lively dance moves, including a tap routine, a spin, and a leap, all with a joyful expression on his face. Throughout the performance, he maintains his polished appearance, complete with a vest, watch chain, and spats, as he channels the spirit of a vaudeville entertainer.]"
},
{
"from": "gpt",
"value": "Good. Basically no problem."
}
] |
seed42_1625 | videos/HRA_part0/1625.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A focused maintenance worker, clad in a dark green jacket over a gray shirt, blue jeans, and brown work boots, kneels on the floor of a clean, organized industrial space. He is attentively replacing batteries in a complex piece of machinery, demonstrating his expertise and care. The scene is illuminated by bright overhead lights, highlighting his meticulous task. Behind him, shelves stocked with tools and equipment provide context for his work environment, emphasizing the importance of regular maintenance in a professional setting.]"
},
{
"from": "gpt",
"value": "Normal. The man's fingers in the video are not fully generated, his facial features are blurry, the proportions of his legs are off, and his feet are not fully generated."
}
] |
seed44_2241 | videos/HRA_part1/2241.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A cheerful farmer, donning a brown shirt and blue overalls, stands in a lush green field under a clear blue sky, smiling as he observes a rabbit. The rabbit, located near the center of the frame, leaps into the air with excitement, its hind legs propelling it forward in a graceful arc. The farmer's joyful expression and the rabbit's energetic jump create a heartwarming scene, capturing the simple pleasures of rural life. The setting is enhanced by the vibrant green grass and the warm sunlight, which bathes the scene in a golden glow.]"
},
{
"from": "gpt",
"value": "Bad. The actions are not continuous and there are sudden changes."
}
] |
seed44_1674 | videos/HRA_part1/1674.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A beaver, typically found in woodland streams, curiously explores a sandy beach near a vibrant beach resort, walking along the shoreline with its distinctive flat tail and brown fur. It passes by colorful umbrellas and sun loungers, a unique sight among the usual beachgoers enjoying the sunny weather. The beaver's presence adds an element of surprise and charm to the bustling beach scene, as it navigates the sandy terrain with its natural grace and agility.]"
},
{
"from": "gpt",
"value": "Good. Basically correct."
}
] |
seed46_1267 | videos/HRA_part2/1267.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A beaver, clad in a brown fedora and matching suspenders over a striped shirt, tap dances on a wooden stage under a spotlight. Behind him, a red velvet curtain provides a dramatic backdrop. He executes a series of lively dance moves, including a tap routine, a spin, and a leap, all with a joyful expression on his face. Throughout the performance, he maintains his polished appearance, complete with a vest, watch chain, and spats, as he channels the spirit of a vaudeville entertainer.]"
},
{
"from": "gpt",
"value": "Bad. The actions are not continuous."
}
] |
seed46_2203 | videos/HRA_part2/2203.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A charismatic man named Jay, donning a gray blazer, black t-shirt, and jeans, stands in a dimly lit room with a warm color tone. Initially, he is seen in profile, gazing into the distance with a reflective look. Jay then turns to face the camera, his expression conveying a sense of determination and strength. Next, he smiles warmly, exuding confidence and approachability. Finally, Jay closes his eyes, appearing to savor a private moment of joy, all while the ambient lighting enhances the intimate atmosphere of the scene.]"
},
{
"from": "gpt",
"value": "Normal. The man's arm generation is missing, and the face generation is blurry."
}
] |
seed44_3042 | videos/HRA_part1/3042.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A distinguished gentleman with salt-and-pepper hair sits in a dimly lit room, wearing a gray blazer and a white shirt with the top button undone. He leans back in a brown leather armchair, holding a book in his hands. The camera captures him from the side, highlighting his profile as he reads intently. Beside him, a wooden table holds a glass of red wine, a vintage lighter, and a half-smoked cigar resting in an ashtray. The atmosphere is cozy and intimate, with the warm glow of a nearby lamp illuminating the scene. Behind the chair, a large bookshelf filled with books adds to the scholarly ambiance. The man's eyes are fixed on the pages, his lips slightly parted, indicating deep concentration as he delves into the story. The setting conveys a sense of leisurely intellectual pursuit, with the elements of wine, cigar, and reading material suggesting a refined evening routine.]"
},
{
"from": "gpt",
"value": "Normal. The man's movements in the video are basically normal, with slight awkwardness."
}
] |
seed46_2101 | videos/HRA_part2/2101.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A handsome, young male actor sprints through a bustling train station, wearing a dark blue coat, khaki pants, and brown shoes, his face a mask of urgency and determination. The vibrant, crowded setting provides a dynamic backdrop, filled with diverse individuals going about their day. His intense gaze and focused posture convey a sense of high stakes, while the surrounding architecture and the moving crowd add depth and realism to the scene. The camera captures his swift movement from multiple angles, highlighting his purposeful stride and the energy emanating from him as he weaves through the throngs of travelers.]"
},
{
"from": "gpt",
"value": "Bad. The character's movements in the video are not smooth."
}
] |
seed42_2128 | videos/HRA_part0/2128.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A casually dressed woman, wearing a beige jacket and dark pants, walks her dog on a lead in a suburban neighborhood. She strolls along the pavement, passing by a row of neatly trimmed hedges and uniform houses with red roofs, creating a sense of domestic tranquility. The dog, a medium-sized brown and white spaniel, trots obediently beside her, its leash held gently in the woman's hand. As they walk, the afternoon sun casts long shadows, highlighting the peacefulness of the scene.]"
},
{
"from": "gpt",
"value": "Bad. The hands of the person in the video are distorted, the legs are distorted, the dog's legs are distorted, and the head is distorted."
}
] |
seed44_1492 | videos/HRA_part1/1492.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A young student, dressed in casual attire of a gray shirt and black shorts, playfully swings a bat at a small pinata hanging from a tree branch in a backyard setting, surrounded by lush greenery. Moments later, the same student is seen enjoying a refreshing swim in a backyard pool, splashing about in a yellow swimsuit, with a diving board visible in the background. The scene captures the essence of youthful fun and leisure activities, transitioning seamlessly from a game of bat-and-ball to a spontaneous dip in the pool.]"
},
{
"from": "gpt",
"value": "Bad. The diving board was not generated in the video."
}
] |
seed46_3023 | videos/HRA_part2/3023.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [An engineer, wearing a dark blue jacket over a gray shirt, and gray pants, stands in a taiga landscape marked by tall coniferous trees and a dense undergrowth of shrubs. Holding a towel in his right hand, he wipes his face, perhaps after a long day of work or physical exertion. The scene is set against the backdrop of the taiga, with the engineer's attire standing out against the natural greens and browns of the forest environment. The act of wiping his face with the towel conveys a sense of relief and respite in the midst of the rugged wilderness.]"
},
{
"from": "gpt",
"value": "Normal. The video did not generate a gray shirt and gray pants."
}
] |
seed42_3205 | videos/HRA_part0/3205.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A small, agile green lizard crawls across a weathered, wooden fence post in a suburban backyard, basking in the warm sunlight that illuminates its textured scales. The fence is partially obscured by lush green foliage, and beyond it, a clear blue sky contrasts with the vibrant greenery, creating a serene and natural atmosphere. The lizard pauses momentarily, alert and attentive, before continuing its crawl along the fence, navigating the rough surface with ease and grace.]"
},
{
"from": "gpt",
"value": "Bad. The lizard's body is twisted, its head is twisted, and its legs are twisted."
}
] |
seed42_2440 | videos/HRA_part0/2440.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A domestic chicken, featuring brown and white feathers, is shown swimming in a small, clear body of water, surrounded by green grass and foliage. It paddles energetically with its feet, creating ripples on the water's surface, while its head remains dry above the waterline. The chicken's unique activity is captured in a series of shots that showcase its unexpected aquatic adventure in a natural, backyard setting.]"
},
{
"from": "gpt",
"value": "Good. Basically no issues."
}
] |
seed44_1672 | videos/HRA_part1/1672.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a futuristic space habitat, a dingo, wearing a small spacesuit, curiously explores the hydroponic garden, gently caressing the leaves of vibrant green plants under the glow of artificial lights designed to mimic sunlight. Nearby, an astronaut, clad in a full spacesuit, tenderly pets the dingo, creating a bond between man, animal, and the pioneering spirit of space agriculture, all against the backdrop of a vast, starlit universe visible through the habitat's clear dome.]"
},
{
"from": "gpt",
"value": "Bad. The generation of the wild dog's head and paws, as well as the astronaut's hand, is problematic in the video."
}
] |
seed46_1516 | videos/HRA_part2/1516.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [An industrious male engineer, wearing a blue button-down shirt and glasses, meticulously irons his clothes in a modestly furnished room with a wooden floor and cream walls. He stands at an ironing board covered with a white cloth, carefully smoothing out wrinkles in a garment. The room is illuminated by natural light from a window behind him, casting a warm glow over the scene. On the wall to his right, a large painting in shades of green adds a splash of color to the otherwise simple space. Beside the window, a desk with a laptop and engineering blueprints hints at his professional life, while a shelf filled with books and a houseplant brings a personal touch to the environment.]"
},
{
"from": "gpt",
"value": "Bad. The iron in hand suddenly disappears, and the fabric in hand suddenly becomes larger and smaller."
}
] |
seed42_2912 | videos/HRA_part0/2912.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a dimly lit diner after hours, a weary waiter, attired in a crisp white shirt, black vest, and striped tie, rests his head on his hands atop a checkered tablecloth. He has removed his shoes and his black socks are visible, adding to the sense of weariness and relaxation. His nametag reads 'Eduardo', and he appears to have fallen into a peaceful slumber, surrounded by the quiet emptiness of the restaurant. A half-drunk cup of coffee sits beside him, a testament to the long night shift he's just completed. The scene is intimate and still, capturing a rare moment of rest for a hardworking individual.]"
},
{
"from": "gpt",
"value": "Bad. The person's hand and arm are distorted in the video."
}
] |
seed42_2656 | videos/HRA_part0/2656.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A film director, dressed in a dark green shirt and beige cargo pants, stands on set with a megaphone in hand, coordinating the action. Nearby, a majestic tiger leaps gracefully over a log, its muscles rippling and eyes intense, demonstrating the power and agility of the big cat. The scene is set in a lush, verdant forest, with sunlight filtering through the leaves, creating a dramatic interplay of light and shadow. The director's focused gaze and the tiger's dynamic leap are captured in high definition, highlighting the intensity and precision required for filming such a scene.]"
},
{
"from": "gpt",
"value": "Good. Basically correct."
}
] |
seed46_1723 | videos/HRA_part2/1723.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A man, dressed in a spacesuit with the helmet removed, floats gracefully in zero gravity inside a space station, illuminated by the ethereal glow of Earth in the background. He gently caresses the inner wall of the station, adorned with complex machinery and wires, while looking out at the breathtaking view of our planet through the large circular window, conveying a sense of wonder and intimacy with the cosmos. The setting captures the essence of human achievement in space exploration, with the man's tender gesture toward the space station's interior wall highlighting a unique connection between humanity and its ventures beyond Earth.]"
},
{
"from": "gpt",
"value": "Normal. The video shows an extra helmet being generated, and the man's limbs are abnormally generated."
}
] |
seed42_25 | videos/HRA_part0/25.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a whimsical scene, a friendly capybara, adorned with a miniature nurse's hat and a stethoscope around its neck, crawls across a colorful, childlike set filled with oversized medical-themed toys. The capybara moves deliberately, as if performing a check-up on a giant teddy bear patient lying on a playful examination table, surrounded by an array of oversized bandages, cotton balls, and a smiling face thermometer, creating a heartwarming and imaginative playtime scenario.]"
},
{
"from": "gpt",
"value": "Normal. The capybara's movements are not smooth."
}
] |
seed46_179 | videos/HRA_part2/179.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a futuristic setting, two children, a boy and a girl, both wearing space suits with helmets under their arms, walk purposefully down a metallic corridor illuminated by blue light. The boy, with short brown hair and a red stripe across his helmet, carries a backpack, while the girl, with long blonde hair tied in a ponytail and a yellow stripe on her helmet, holds a small, mysterious box. They navigate through a door marked with alien symbols, indicating their arrival at a space diplomatic mission. Inside, they meet a diverse group of extraterrestrial beings, all gathered around a large holographic map, creating an atmosphere of intergalactic diplomacy. The children, standing tall among these alien delegates, represent humanity on this important interstellar journey.]"
},
{
"from": "gpt",
"value": "Good. Basically correct."
}
] |
seed46_2445 | videos/HRA_part2/2445.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a serene grassland setting under a clear blue sky, a police officer, clad in a standard uniform with a badge and gun holster, kneels by a small, clear pond. He watches intently as a group of children joyfully splash and swim in the water, their laughter echoing across the open landscape. The officer's presence is both protective and relaxed, creating a harmonious scene where law enforcement and community enjoyment coexist peacefully amidst the natural beauty of the vast, green grasslands.]"
},
{
"from": "gpt",
"value": "Good. Basically correct."
}
] |
seed44_1580 | videos/HRA_part1/1580.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A middle-aged janitor, wearing a yellow safety vest over a gray shirt, blue jeans, and brown shoes, climbs a tall, metal, industrial ladder in a large warehouse filled with boxes and equipment. He holds a tool in his right hand and reaches up with his left, the afternoon sun shining through the high windows, casting a warm glow over the dusty space. The ladder leans against a tall shelf stacked with supplies, and the man's focus is intense as he ascends, his brows furrowed in determination. Behind him, the vastness of the warehouse stretches out, with rows of shelves disappearing into the distance, emphasizing the scale of his task.]"
},
{
"from": "gpt",
"value": "Bad. No brown shoes were generated, the ladder position is incorrect, and the character's hands are in the wrong state."
}
] |
seed44_2610 | videos/HRA_part1/2610.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A well-dressed lawyer, wearing a gray suit with a blue shirt and striped tie, stands in a modern office space. He is attentively checking a smoke detector mounted on the ceiling, exhibiting a careful and professional demeanor. The office features a large window offering views of a city skyline, a desk with a laptop and documents, and a bookshelf filled with law books, creating a setting that blends professionalism with safety awareness. The lawyer's actions highlight the importance of routine safety checks in a workplace environment.]"
},
{
"from": "gpt",
"value": "Bad. The lawyer's movements in the video are very unsmooth and uncoordinated."
}
] |
seed44_1949 | videos/HRA_part1/1949.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A man, appearing to be in his thirties, with short brown hair and wearing a grey t-shirt and black pants, is shown in a close-up shot. He exhibits signs of severe trembling, with beads of sweat visible on his forehead, indicating intense physical or emotional distress. The shaky movement is the focal point, capturing the viewer's attention and conveying a sense of urgency and concern. The background is blurred, keeping the emphasis solely on the man's shaking figure, which fills the screen, making his distress the central theme of the scene.]"
},
{
"from": "gpt",
"value": "Good. No obvious problems."
}
] |
seed42_2501 | videos/HRA_part0/2501.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a dimly lit cave, a police officer, wearing a uniform with a badge and a utility belt equipped with various tools, including a flashlight, gently caresses the rough stone wall, seemingly searching for hidden clues or evidence. The cave's textured walls, illuminated by the officer's flashlight, provide a dramatic backdrop for this intense moment of investigation. The officer's focused gaze and careful touch convey a sense of determination and meticulousness in uncovering secrets within the cave's depths.]"
},
{
"from": "gpt",
"value": "Normal. The officer's action of touching the stone wall in the video is not obvious, but the movement is basically normal."
}
] |
seed44_1056 | videos/HRA_part1/1056.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A mature silverback gorilla sits in a lush, green environment, gently caressing a smaller gorilla with his large, expressive hands, showing tenderness and care. The silverback, with his powerful build and wise eyes, tenderly strokes the younger gorilla's back, creating a heartwarming scene of connection and affection within the primate family. This intimate moment is framed by the verdant foliage of their natural habitat, highlighting the emotional bond between these two creatures.]"
},
{
"from": "gpt",
"value": "Bad. The larger silverback gorilla's arms are twisted and deformed, and the smaller silverback gorilla's head is deformed."
}
] |
seed46_3032 | videos/HRA_part2/3032.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a bustling cityscape dominated by towering office buildings, a compassionate doctor, wearing a crisp white lab coat, leans in to gently caress the hand of a seated patient. The patient, dressed in casual attire, looks up at the doctor with gratitude, finding solace in the medical professional's empathetic touch. The scene is set inside a modern, well-lit clinic, with sleek furniture and medical equipment in the background, emphasizing the doctor's dedication to providing care in a professional environment.]"
},
{
"from": "gpt",
"value": "Bad. In the video, the doctor can be clearly seen touching the patient's hand, and the movements are very discontinuous and uncoordinated."
}
] |
seed44_2017 | videos/HRA_part1/2017.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a vast, sun-drenched savanna, two graceful antelopes engage in a tender moment of affection, their brown coats glistening under the golden light. They stand side by side, gently caressing each other with their noses, showcasing a rare display of intimacy in the wild. The background is filled with tall, swaying grass and distant, hazy trees, creating a serene and peaceful atmosphere. Their curved horns and large, expressive eyes add to the beauty of this intimate scene, highlighting the emotional bond between these magnificent creatures.]"
},
{
"from": "gpt",
"value": "Bad. The antelope's head is twisted, neck is twisted, and legs are twisted."
}
] |
seed46_1920 | videos/HRA_part2/1920.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A tiny, adorable quail chick, no bigger than a thumb, stands on a wooden surface, its downy feathers a mix of brown and cream. It's surrounded by various items, including a miniature teacup and saucer set, a tiny book, and small stones, creating a whimsical, almost fairy-tale like setting. The chick looks curious and alert, pecking at the objects around it. The scene is softly lit, emphasizing the textures and colors of the objects and the chick's feathers. The backdrop is a blurred green, giving the impression of an outdoor setting, perhaps a garden or park, further enhancing the storybook quality of the scene.]"
},
{
"from": "gpt",
"value": "Good. Basically no problem."
}
] |
seed42_2620 | videos/HRA_part0/2620.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A police officer, wearing a navy blue uniform and badge, meticulously organizes his home garage, filled with various tools and equipment. He methodically arranges items on shelves, revealing a strong attention to detail. The garage space, well-lit and clean, reflects his disciplined nature. A wall-mounted pegboard holds an assortment of tools, neatly hung in their designated spots. Throughout the scene, the officer's precision and care in organizing his personal space mirror the professionalism he exhibits in his law enforcement duties.]"
},
{
"from": "gpt",
"value": "Good. Basically correct."
}
] |
seed46_2069 | videos/HRA_part2/2069.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A farmer wearing a red shirt and brown overalls strolls through a vast, lush green field, holding a walking stick as he leisurely promenades. Suddenly, he leaps into the air, his hat flying off, with a wide smile on his face, revealing sheer joy and freedom. The farmer continues his playful jumping spree amidst the vibrant scenery, surrounded by endless rows of corn stalks swaying gently in the breeze. As he lands, he resumes his peaceful walk, stick in hand, with the sun casting a warm glow over the fertile land, creating a sense of harmony and contentment.]"
},
{
"from": "gpt",
"value": "Bad. The character in the video has a twisted body, distorted face, twisted arms, twisted hands, and twisted legs."
}
] |
seed44_2278 | videos/HRA_part1/2278.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A vibrant musician, donning a yellow shirt, catches a flying drumstick during an energetic performance under the spotlight. The scene transitions to show the artist passionately singing into the microphone, the crowd's glow sticks illuminating the dark venue. Next, the musician is seen catching a guitar pick from another band member, their connection highlighted by the stage lights. Finally, the artist reaches out to catch a flower thrown from the audience, a gesture symbolizing the bond between performer and fans, all set against a backdrop of colorful stage lighting.]"
},
{
"from": "gpt",
"value": "Bad. The musician mentioned in the caption taking the guitar pick from another band member does not appear in the video."
}
] |
seed42_3150 | videos/HRA_part0/3150.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a surprising yet tender scene, a majestic male lion gently nuzzles and caresses a smaller lion, possibly his cub, against the backdrop of a lush, green savanna during golden hour. The adult lion, with its voluminous tawny mane, displays affection, while the cub, sporting a lighter coat and smaller stature, responds with innocent playfulness. The heartwarming interaction unfolds under a vast, cloud-dappled sky, with scattered acacia trees and distant mountains completing the serene African landscape. This intimate moment captures the rare display of tenderness and bonding within the lion family, far from the typical fierce and powerful image associated with these animals.]"
},
{
"from": "gpt",
"value": "Normal. The lion's face is blurry."
}
] |
seed42_2315 | videos/HRA_part0/2315.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A majestic lion, king of the beasts, is seen in the early morning, wiping its face with its paw in the grasslands of Africa. The lion, with a full, tawny mane, is lying down, using its right paw to wipe its face, perhaps after a recent meal or simply as a grooming habit. The grassy savanna stretches out behind it, dotted with acacia trees, providing a natural habitat for this powerful creature. The sun, still low in the sky, casts a warm glow over the scene, highlighting the lion's golden fur and the textures of the grass.]"
},
{
"from": "gpt",
"value": "Bad. The lion wiping its face with its right paw, mentioned in the caption, does not appear in the video."
}
] |
seed44_1237 | videos/HRA_part1/1237.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A distinguished film director, donning a classic tweed suit, sits on a wooden chair on a meticulously arranged movie set, gently caressing a fluffy white rabbit with a careful, nurturing touch. The director, with a neatly trimmed beard, looks intently at the rabbit, showing a softer side amidst the industrial backdrop of the set. As the scene progresses, the director continues to pet the rabbit affectionately, the camera capturing every tender gesture in high definition, highlighting the contrast between the tough world of filmmaking and the director's gentle interaction with the small, vulnerable creature.]"
},
{
"from": "gpt",
"value": "Normal. The white rabbit in the video was generated abnormally, and the director's hand was generated with disproportionate proportions."
}
] |
seed44_1214 | videos/HRA_part1/1214.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A whimsical scene unfolds where a hedgehog, adorned with miniature headphones and a tiny guitar strapped to its back, embarks on an adventurous journey across a lush, green backyard. It navigates through grassy terrains, darting between vibrant flowers and undergrowth with a sense of purpose. The hedgehog's spines glisten under the golden sunlight, adding a touch of magic to the scene. As it runs, the camera captures the hedgehog from various angles, highlighting its determination and the enchanting world it inhabits. The setting is reminiscent of a fairy tale garden, complete with a wooden fence in the background, framing the hedgehog's spirited escapade.]"
},
{
"from": "gpt",
"value": "Normal. The hedgehog's back is missing a guitar."
}
] |
seed46_1303 | videos/HRA_part2/1303.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A North American porcupine, distinguished by its brown and white quills, runs through a dense forest environment, navigating between moss-covered trees. The porcupine, with its quills standing out prominently, dashes across the screen from right to left, showcasing its agility. Its facial features, including small eyes and a pointed nose, are clearly visible as it moves quickly, demonstrating the animal's unexpected speed and grace in its natural habitat.]"
},
{
"from": "gpt",
"value": "Bad. The porcupine in the video has a twisted body, a distorted face, and twisted legs."
}
] |
seed42_2686 | videos/HRA_part0/2686.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A middle-aged female teacher, wearing a dark blue blazer over a white blouse and carrying a black handbag, walks purposefully down a wide hallway with high ceilings, resembling the interior of a university building. The walls are adorned with educational posters, and the floors are covered with brown carpeting. As she strides forward, the scene subtly transitions to reveal a large window behind her, displaying a breathtaking view of Neptune's turbulent atmosphere, filled with vivid blue and green clouds swirling majestically. This unique setting, combining the familiar environment of a school with the awe-inspiring backdrop of a distant planet, creates a captivating visual narrative.]"
},
{
"from": "gpt",
"value": "Normal. The video clearly shows the female teacher walking, with basically normal movements, but there is some lack of coordination."
}
] |
seed42_2273 | videos/HRA_part0/2273.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a futuristic setting illuminated by vibrant neon lights, a diligent worker clad in a gray jumpsuit and helmet stands on a sleek, anti-gravity vehicle. The vehicle, featuring a metallic finish and aerodynamic design, hovers above a clean, white platform. The worker is engaged in wiping the vehicle's surface with a cloth, ensuring its spotless condition. Behind the scene, a vast hangar filled with advanced technology and equipment can be seen, emphasizing the high-tech environment in which the worker operates. The worker's focused expression and meticulous attention to detail highlight their commitment to maintaining the anti-gravity vehicle's pristine state.]"
},
{
"from": "gpt",
"value": "Normal. The person's arm in the video is twisted, the hand is twisted, and the cloth in the hand is twisted."
}
] |
seed44_2917 | videos/HRA_part1/2917.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A small, fluffy hamster sits upright on its hind legs, engaging in a grooming session within its cozy cage. The hamster meticulously wipes its face with its front paws, demonstrating a natural behavior typical of these animals. Its movements are deliberate and focused, highlighting the creature's self-care routine in a habitat filled with soft bedding and a water bottle visible in the background. The hamster's fur is a mix of brown and white, adding to the visual appeal of the scene as it diligently cleans itself.]"
},
{
"from": "gpt",
"value": "Bad. The hamster's face rubbing in the video is very unsmooth, and the movements are uncoordinated."
}
] |
seed44_109 | videos/HRA_part1/109.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a surreal, futuristic scene, a police officer, clad in a high-tech, silver uniform with a clear, bubble-like helmet, leaps gracefully over a rocky terrain under the Martian sky. An ibex, a type of wild goat, stands majestically nearby, its silhouette outlined against the alien landscape. The officer's powerful jump is captured mid-air, highlighting the reduced gravity of Mars. The constellations and stars above are vivid and numerous, creating a breathtaking backdrop for this extraordinary moment of stargazing from the red planet.]"
},
{
"from": "gpt",
"value": "Bad. The police officer's movements are uncoordinated and discontinuous, with sudden changes."
}
] |
seed46_3062 | videos/HRA_part2/3062.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a dimly lit room, a thoughtful writer sits at a cluttered desk, typing away on a laptop with determination. Suddenly, the scene focuses on a small vole, its fur a mix of brown and gray, crawling out from beneath the wooden desk. The rodent explores its surroundings, nibbling on a pencil and climbing onto a pile of books, all while the writer continues to work, seemingly unaware of the tiny creature's presence.]"
},
{
"from": "gpt",
"value": "Good. No obvious issues."
}
] |
seed46_2550 | videos/HRA_part2/2550.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A janitor, clad in a yellow vest over a dark jacket and jeans, walks purposefully down a hospital corridor. The hall is lined with evenly spaced white doors and illuminated by overhead lights, casting a warm glow. The man holds a mop bucket in one hand, its handle adorned with a colorful cloth hanging loosely. He moves with determination, his work boots echoing softly on the linoleum floor, creating a sense of routine and responsibility in the clean, institutional environment.]"
},
{
"from": "gpt",
"value": "Normal. The cleaner's hands and feet in the video have generation issues, and the mop bucket is deformed."
}
] |
seed42_198 | videos/HRA_part0/198.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A chef, donning a classic white uniform with a red bandana around his neck, walks through a lush, green forest, his eyes fixed ahead with determination. Accompanied by a small, bushy-tailed squirrel perched on his shoulder, they navigate the woodland together. The chef carries a large basket, presumably filled with foraged ingredients, while the squirrel clutches an acorn in its paws, both on a mission in the vibrant, sun-dappled setting.]"
},
{
"from": "gpt",
"value": "Good. No obvious problems."
}
] |
seed46_2312 | videos/HRA_part2/2312.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A professional male chef, donning a white chef's uniform with black stripes and a matching cap, meticulously washes dishes in a commercial kitchen. The room is illuminated by bright overhead lights, casting a glow on the stainless steel sink and countertops. Steam rises from the hot water, creating a misty atmosphere. The chef's sleeves are rolled up to his elbows, revealing tattoos, as he focuses intently on cleaning each plate with precision and care. In the background, the bustling kitchen is visible, with other chefs preparing meals and the sounds of clattering pots and sizzling pans adding to the lively ambiance.]"
},
{
"from": "gpt",
"value": "Bad. The man's hand and arm appear deformed and mutated, and the plate in his hand also becomes deformed."
}
] |
seed46_2800 | videos/HRA_part2/2800.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A majestic bald eagle soars through the crisp, blue sky, its powerful wings outstretched against a backdrop of puffy white clouds. The eagle's piercing yellow eyes and sharp, curved beak are sharply detailed. It effortlessly glides over a vast, serene lake, surrounded by lush, green coniferous trees and a rugged mountain range. The eagle's regal presence is highlighted by the sun's golden rays filtering through the clouds, casting a warm glow on the landscape below. The bird's brown and white feathers ruffle in the wind as it continues its graceful flight, embodying freedom and strength in the wild.]"
},
{
"from": "gpt",
"value": "Good. No obvious problems."
}
] |
seed44_111 | videos/HRA_part1/111.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A police officer, clad in a navy uniform with gold accents and a peaked cap, runs purposefully across a grand hall within a futuristic space station. The hall is lined with flags from various nations and interspersed with abstract art pieces, all illuminated by the soft glow of recessed lighting. In the distance, a large, circular window reveals the vastness of space and the distant twinkle of stars. The officer's run suggests urgency, possibly related to security matters at the ongoing political summit, where delegates from Earth and various space colonies gather to discuss interstellar policy. The polished floor reflects the officer's movement, emphasizing the gravity of the situation in this otherwise serene and majestic setting.]"
},
{
"from": "gpt",
"value": "Normal. The character's head is deformed, and the arms are twisted."
}
] |
seed42_146 | videos/HRA_part0/146.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a vibrant, sun-dappled deciduous forest during autumn, several joyful people gather for an impromptu dance party. The scene opens with a close-up of a smiling woman in a floral dress, her arms raised as she twirls under the golden leaves. Next, a group of friends, dressed in casual fall attire like sweaters and jackets, laugh and dance together, surrounded by tall trees and colorful foliage. The camera then captures a couple slow dancing romantically, their embrace tender and intimate, with the forest's warm hues providing a romantic backdrop. Lastly, the entire group comes together, forming a circle as they dance energetically, their laughter and the rustling leaves blending into a symphony of happiness in the heart of the forest.]"
},
{
"from": "gpt",
"value": "Good. Basically correct."
}
] |
seed42_2547 | videos/HRA_part0/2547.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A janitor, clad in a yellow safety vest and holding a mop, stands near a large window overlooking a kangaroo casually hopping through the grounds of a power plant complex, set against a backdrop of industrial cooling towers. The scene transitions to show the janitor, now in swimming trunks, diving into a clear, blue pool, while a kangaroo watches from the edge, all under the watchful gaze of the towering power plant structures in the distance. The final shot captures the unexpected duo, the janitor and the kangaroo, enjoying a peaceful moment together by the poolside, with the power plant's silhouette framing the serene scene.]"
},
{
"from": "gpt",
"value": "Bad. The kangaroo was not generated in the swimming pool scene."
}
] |
seed42_258 | videos/HRA_part0/258.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A compassionate female veterinarian, wearing a white lab coat and blue scrubs, gently feeds a small dog from a spoon, her caring nature evident in her warm smile. The clinic room is clean and modern, with stainless steel equipment and a large window providing natural light. Behind her, a poster about pet nutrition hangs on the pale green wall. On a nearby table, various pet food samples are neatly arranged, emphasizing the importance of proper animal care. Throughout the scene, the vet's loving interaction with the dog highlights her dedication to animal welfare.]"
},
{
"from": "gpt",
"value": "Good. Basically no issues."
}
] |
seed42_1445 | videos/HRA_part0/1445.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A magnificent cheetah, adorned with vibrant orange beads and feathers woven into its fur, dances gracefully in the dappled sunlight of a lush savanna. Its body moves fluidly to an unheard beat, each leap and twist a display of raw power and elegance. Surrounded by tall grass swaying gently in the breeze, the cheetah's every motion is captured in high definition, from the flex of its muscles to the gleam in its eyes, embodying the spirit of the wild in a captivating dance.]"
},
{
"from": "gpt",
"value": "Bad. The cheetah's body is twisted, its paws are twisted, its face is distorted, and its legs are twisted."
}
] |
seed46_1734 | videos/HRA_part2/1734.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A nocturnal porcupine, with its distinctive quills standing erect, is captured mid-jump in a dimly lit forest environment, surrounded by lush greenery. The quills, which appear sharp and menacing, contrast against the soft foliage of the underbrush. The porcupine's muscular hind legs propel it into the air, showcasing its surprising agility, while its front paws reach forward, ready to grasp onto a tree trunk or branch. The scene is set against a backdrop of tall trees and the ambient sounds of night creatures, highlighting the porcupine's natural habitat.]"
},
{
"from": "gpt",
"value": "Normal. The porcupine's body is twisted in the video."
}
] |
seed42_3084 | videos/HRA_part0/3084.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A middle-aged male chef, wearing a traditional white chef's uniform and black pants, stands outdoors in a garden, holding a shovel. He is surrounded by various plants, including leafy greens, carrots, and herbs like rosemary, thyme, and sage. The chef has short brown hair and a neatly trimmed beard, and he is digging into the rich soil with determination, perhaps searching for fresh ingredients for his culinary creations. The scene is set against a backdrop of wooden crates and greenery, with a hint of sunlight filtering through the trees, casting dappled shadows on the chef and his gardening endeavor.]"
},
{
"from": "gpt",
"value": "Good. Basically no issues."
}
] |
seed42_233 | videos/HRA_part0/233.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A rugged farmer, donning a plaid shirt, denim overalls, and a worn baseball cap, stands in a modern kitchen. He methodically unloads a sleek, stainless steel dishwasher, handling each plate and utensil with unexpected gentleness. The contrast between his calloused hands and the shiny, clean dishes highlights the unusual domestic task for someone more accustomed to farm chores. The kitchen is spacious and well-lit, featuring granite countertops and contemporary appliances, offering a glimpse into the farmer's home life beyond the fields.]"
},
{
"from": "gpt",
"value": "Bad. The character's movements are not smooth."
}
] |
seed44_3166 | videos/HRA_part1/3166.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a heartwarming scene, a compassionate female doctor, wearing a white lab coat and blue surgical mask, gently caresses an injured elephant's trunk. They stand under a colossal, weathered stone bridge, its arches towering above them, providing a dramatic backdrop. The elephant, adorned with vibrant red and yellow decorations, hints at cultural significance, while the bridge's ancient stonework contrasts with the modern attire of the doctor, highlighting the juxtaposition of old and new. The tender interaction between the two, set against the imposing structure of the bridge, conveys a message of care and connection across species and time.]"
},
{
"from": "gpt",
"value": "Normal. The action of the female doctor touching the elephant's trunk is slightly unsmooth."
}
] |
seed42_2456 | videos/HRA_part0/2456.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A majestic bald eagle soars through a clear blue sky, its wings spread wide, revealing its distinctive white-feathered head and brown body. It glides effortlessly over a vast, serene landscape, comprised of lush green forests and tranquil lakes, with distant mountain peaks forming a stunning backdrop. The eagle's piercing yellow eyes remain alert, scanning the terrain below for any sign of prey, as it embodies the spirit of freedom and strength in the untamed wilderness.]"
},
{
"from": "gpt",
"value": "Good. No obvious issues."
}
] |
seed46_2341 | videos/HRA_part2/2341.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A white hare, its fur thick and plush, crawls through a dense thicket of snow-laden pine trees, the texture of its coat contrasting starkly with the icy landscape. The hare's movements are cautious and deliberate, its eyes alert, as it navigates the snowy terrain in search of food. The scene is set against a backdrop of a serene winter forest, the hare's crawling motion emphasized by the slow, methodical pace, creating a sense of intimacy with the animal in its natural habitat.]"
},
{
"from": "gpt",
"value": "Normal. The rabbit's movements in the video are basically normal, but there is slight uncoordination."
}
] |
seed46_2920 | videos/HRA_part2/2920.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A waiter, donning a crisp white shirt, black trousers, and a waistcoat, unexpectedly finds himself fixing a leaky faucet in a vintage-styled restroom. With a wrench in hand, he leans over the basin, his face concentrated as he works under the sink. The scene captures the unusual sight of the waiter, his attire contrasting with the task at hand, as he tightens the faucet, the water now streaming smoothly. The room is adorned with antique fixtures and a classic mirror above the basin, adding to the unique scenario of a waiter turned handyman.]"
},
{
"from": "gpt",
"value": "Bad. The doctor's face and hands are distorted in the video."
}
] |
seed44_2934 | videos/HRA_part1/2934.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A well-dressed lawyer, wearing a dark suit and carrying a leather briefcase, walks with purpose across a dimly lit, gritty urban environment. Simultaneously, a CGI weasel, also dressed in a miniature lawyer's outfit complete with a bow tie and a tiny briefcase, scurries alongside the human lawyer. Both characters navigate through the cityscape, which features brick walls, fire escapes, and a sense of the urban jungle. The juxtaposition of the human lawyer and the anthropomorphized weasel lawyer adds a layer of whimsy and metaphor to the scene, suggesting themes of cunning and legal maneuvering.]"
},
{
"from": "gpt",
"value": "Good. No obvious problems."
}
] |
seed46_2911 | videos/HRA_part2/2911.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A young student, dressed in a casual outfit of a gray sweater and blue jeans, sits cross-legged on a wooden floor inside a cozy room, surrounded by bookshelves. In front of her is a small, curious hedgehog, slowly crawling across a white towel, exploring its immediate environment with keen interest. The student watches the hedgehog intently, her face a mix of fascination and care, as she gently interacts with the animal, creating a heartwarming scene of connection between human and nature indoors.]"
},
{
"from": "gpt",
"value": "Normal. The hedgehog's movements in the video are slightly unsmooth."
}
] |
seed44_130 | videos/HRA_part1/130.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [An engineer, wearing a blue plaid shirt, khaki pants, and glasses, sits in a wooden chair in a cozy living room filled with natural light. They hold a crochet hook and yarn, working on a colorful blanket with a complex geometric pattern. The space around them is cluttered with engineering books, blueprints, and tools, blending the worlds of precision and creativity. The camera captures the intricate process, focusing on their hands as they expertly loop the yarn, creating a vivid contrast between the technical skills of engineering and the artistic craft of crocheting.]"
},
{
"from": "gpt",
"value": "Bad. The video did not generate glasses, crochet hooks, or colorful yarn."
}
] |
seed44_3172 | videos/HRA_part1/3172.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A distinguished lawyer, donning a navy suit with a crisp white shirt and a patterned tie, sits at an opulent wooden desk. The room is richly decorated with leather-bound books and classic art pieces. He lifts a fork filled with a gourmet meal to his mouth, savoring each bite. The table is set with fine china and silverware, a half-empty glass of red wine standing beside a nearly finished plate of food, indicating a moment of leisure in his busy schedule. Soft ambient light from a nearby lamp illuminates the scene, casting warm shadows across the polished wood surface.]"
},
{
"from": "gpt",
"value": "Bad. The character's face is distorted, hands are distorted, and the plate is distorted."
}
] |
seed42_1255 | videos/HRA_part0/1255.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a surreal space setting filled with stars and galaxies, a beautifully dressed couple exchanges vows during an extraordinary wedding ceremony. A majestic gazelle, adorned with a golden collar and standing on a floating platform nearby, watches the event with calm interest. The bride, in a flowing white gown, and the groom, in a tailored black suit, hold hands tenderly, their faces reflecting deep love and commitment. As they caress each other's faces, the gazelle gently nuzzles the bride's arm, adding a touch of magic and wonder to this celestial union.]"
},
{
"from": "gpt",
"value": "Normal. The character's movements are not smooth."
}
] |
seed44_1315 | videos/HRA_part1/1315.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A student, dressed in a yellow t-shirt and blue jeans, is playfully chasing a juvenile rhinoceros across a vast, grassy savannah under a bright blue sky. The action is captured from multiple angles, first showing the student's enthusiastic pursuit, followed by a shot of the rhino's surprised reaction as it turns its head back. Next, the scene focuses on the student's determined facial expression, highlighting the joy and energy of the chase. Finally, a side view captures the dynamic duo running in unison, the student's arms waving wildly and the rhinoceros' ears flapping, both enjoying the spontaneous adventure in the expansive landscape.]"
},
{
"from": "gpt",
"value": "Bad. The character's movements are not smooth, and the rhino's movements are not smooth."
}
] |
seed46_2638 | videos/HRA_part2/2638.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a scene directed with artistic flair, a bobcat, known for its secretive nature, emerges from the dense undergrowth of a temperate deciduous forest. The sunlight filters through the canopy, casting dappled shadows on the ground as the bobcat approaches a serene, crystal-clear pool. With a sudden burst of energy, it leaps into the water, its powerful muscles propelling it forward as it swims gracefully beneath the surface. Bubbles rise around it, creating an enchanting underwater ballet, before it resurfaces, shaking droplets from its fur, and returns to the safety of the forest edge, all captured in stunning high-definition detail.]"
},
{
"from": "gpt",
"value": "Good. Basically correct."
}
] |
seed44_1943 | videos/HRA_part1/1943.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A small, vibrant green lizard with intricate patterns adorns its skin, positioned on a dry, cracked earth surface under a bright, cloud-dappled sky. It appears to dance, flicking its tongue and swaying its body in a manner that suggests movement to an unseen beat. The lizard's movements are fluid and energetic, its tail swishing and its limbs moving gracefully as if it's performing a choreographed routine. Despite the arid environment, the lizard's dance brings a sense of lively energy to the scene.]"
},
{
"from": "gpt",
"value": "Normal. The sky full of clouds mentioned in the caption does not appear in the video."
}
] |
seed42_1505 | videos/HRA_part0/1505.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A small, plump quail with brown and black speckled feathers stands prominently in the foreground, nestled among lush green foliage that includes leaves, ferns, and tall grasses. It faces left, its head turned slightly toward the camera, revealing its distinctive facial features, including a small beak and alert eyes. The bird's textured feathers are intricately detailed, and its legs are visible as it perches on the forest floor, surrounded by the vibrant flora.]"
},
{
"from": "gpt",
"value": "Good. No obvious problems."
}
] |
seed42_1440 | videos/HRA_part0/1440.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A charismatic individual named Jay, donning a dark blue cap, aviator sunglasses, and a gray shirt, is captured in high-definition. Initially, he is seen against a backdrop of lush greenery, standing tall with his hands on his hips, exuding confidence. The scene then focuses on his face, highlighting his stylish appearance and a slight smile. Next, Jay is shown from behind, gazing out at a stunning mountainous landscape, conveying a sense of adventure. Finally, he raises his right arm, pointing at something exciting in the distance, his silhouette framed by the natural beauty around him, encapsulating the essence of an outdoor enthusiast.]"
},
{
"from": "gpt",
"value": "Good. Basically no problem."
}
] |
seed44_2696 | videos/HRA_part1/2696.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A North American beaver, known for its brown fur and flat, paddle-like tail, crawls along the muddy bank of a tranquil river. It pauses for a moment, gnawing on a piece of wood with its sharp incisors, before continuing its slow, deliberate crawl. Its small, rounded ears and half-closed eyes add to the relaxed atmosphere of the scene. The beaver's movement is unhurried, as it navigates the earthy terrain near the water's edge, occasionally pausing to survey its surroundings with a curious tilt of its head.]"
},
{
"from": "gpt",
"value": "Normal. The beaver in the video is mostly not moving."
}
] |
seed44_1245 | videos/HRA_part1/1245.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A cheerful musician, adorned in a red apron over a casual black t-shirt and dark jeans, skillfully chops fresh vegetables on a wooden cutting board in a vibrant kitchen filled with natural light. The room is alive with warm colors and modern decor, featuring a large window that lets in the afternoon sun. The musician, with short brown hair and a friendly smile, gracefully moves about the space, stirring a sizzling pan on a high-end stove, the aroma of a delicious meal blending with the upbeat tunes playing softly in the background. Moments later, they taste-test the simmering sauce, adjusting the seasoning with a practiced hand, all while the inviting glow of the kitchen enhances the cozy atmosphere.]"
},
{
"from": "gpt",
"value": "Bad. The knife in the person's hand suddenly turned into a shovel, and the person's hand became deformed."
}
] |
seed46_2227 | videos/HRA_part2/2227.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A vibrant woman with a charming smile, wearing a casual graphic T-shirt and dark jeans, is laughing joyously. Her laughter is contagious as she throws her head back, her wavy brown hair cascading down her shoulders. The scene is set in a cozy living room with wooden floors and a white couch, adding to the warm and intimate atmosphere. Her joyful expression, complete with closed eyes and a wide grin, radiates happiness, making the viewer feel like they're sharing in her delight.]"
},
{
"from": "gpt",
"value": "Bad. The character's movements in the video are not smooth."
}
] |
seed42_1400 | videos/HRA_part0/1400.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A European badger, with its distinct black and white striped face and grey fur, is seen in a natural woodland setting. It stands on its hind legs, using its front paws to wipe its face meticulously, cleaning itself in the early morning light. The action is repeated several times, emphasizing the creature's grooming habits, all while surrounded by lush green foliage and undergrowth, giving a sense of the animal's natural habitat.]"
},
{
"from": "gpt",
"value": "Normal. The hands and mouth of the European badger in the video are not fully generated."
}
] |
seed46_1647 | videos/HRA_part2/1647.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A vast, open ocean landscape fills the screen, with crystal clear water revealing a sandy bottom far below. Sunlight filters through the surface, casting dappled light on the scene. Suddenly, a magnificent manta ray enters the frame, its massive wings spanning wide as it glides effortlessly through the water. The creature is a stunning sight, its dark top and white underbelly contrasting beautifully against the blue surroundings. It performs an elegant somersault, showcasing its agility and grace, before continuing on its serene journey across the vast underwater expanse.]"
},
{
"from": "gpt",
"value": "Bad. The manta ray's body is generated disharmoniously, and the head is not generated."
}
] |
seed44_1106 | videos/HRA_part1/1106.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A majestic red deer leaps over a wooden fence in a lush, green field, its powerful legs and large frame displayed against a backdrop of trees and rolling hills. The deer, with an impressive set of antlers, is captured mid-jump, showing its agility and grace as it clears the barrier. The scene is set in a natural, countryside environment, with the deer's movement frozen in time, highlighting the beauty and strength of this wild animal.]"
},
{
"from": "gpt",
"value": "Bad. The red deer's running in the video is not smooth, and the movements are uncoordinated."
}
] |
seed42_2325 | videos/HRA_part0/2325.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A distinguished doctor, wearing a white lab coat over a button-down shirt and khaki pants, rearranges the furniture in a modern living room. He first adjusts a wooden table, pushing it closer to a brown leather sofa. Next, he moves a gray armchair, placing it at an angle near the sofa, creating a cozy seating arrangement. The doctor then fine-tunes the position of a glass coffee table, centering it between the sofa and armchair. Lastly, he adds a touch of greenery by positioning a potted plant next to the armchair, completing the refreshed layout of the room.]"
},
{
"from": "gpt",
"value": "Bad. The character's face, hands, and arms are distorted in the video."
}
] |
seed44_1686 | videos/HRA_part1/1686.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A fluffy, baby panda with distinctive black and white fur crawls across a lush, green forest floor, surrounded by vibrant vegetation and bamboo shoots. It moves with a curious, almost playful demeanor, its tiny paws gripping the mossy ground. The scene is filled with natural light filtering through the dense canopy above, casting dappled shadows on the panda as it explores its environment.]"
},
{
"from": "gpt",
"value": "Good. No obvious problems."
}
] |
seed46_2962 | videos/HRA_part2/2962.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A male worker, wearing a gray T-shirt with a logo, blue jeans, and a dark cap, assembles furniture in a bright room. He focuses on tightening screws on a wooden dresser using a cordless drill, demonstrating his expertise. The room is filled with natural light, illuminating the task at hand. Moments later, he expertly aligns the dresser drawers, ensuring they fit perfectly. Lastly, he stands back to admire his completed work, a fully assembled dresser standing against a white wall, with the sunlight highlighting the wood grain and the worker's satisfaction in his craftsmanship.]"
},
{
"from": "gpt",
"value": "Bad. The man's head generation in the video has problems, his fingers are deformed, and the cordless drill in his hand suddenly disappears in the middle frame."
}
] |
seed46_2908 | videos/HRA_part2/2908.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A distinguished film director, clad in a black puffer jacket, dark jeans, and black boots, stands on a vibrant movie set. He momentarily closes his eyes, blinking, before opening them wide again to maintain focus. As he blinks once more, the overhead lights reflect off his glasses, hinting at the intense concentration required for his role. Despite the bustling activity around him, he remains centered, pausing briefly to blink, his attire and the lively set creating a contrast between the calm director and the dynamic filmmaking environment.]"
},
{
"from": "gpt",
"value": "Good. Basically correct."
}
] |
seed42_2681 | videos/HRA_part0/2681.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [An engineer with short brown hair, wearing safety glasses, a gray shirt, and khaki pants, stands in a large industrial space filled with machinery. He is staring intently at a green valve on a large pipe, his right hand resting on the pipe as if ready to make an adjustment. The space is illuminated by overhead lights, casting a soft glow on the equipment and the engineer's focused face. Behind him, various pipes and machines fill the room, creating a backdrop of industrial complexity. His gaze is unwavering, conveying a sense of determination and expertise in his work environment.]"
},
{
"from": "gpt",
"value": "Good. Basically correct."
}
] |
seed42_1493 | videos/HRA_part0/1493.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A young student, dressed in a vibrant red swim cap and goggles, is seen playfully interacting with a curious fox at the edge of a serene, crystal-clear pool. The scene opens with the student, wearing a blue swimsuit, diving into the water with a graceful splash, while the fox watches attentively from the side. As the student swims, the fox paces alongside the pool, its interest piqued by the water. Later, the student emerges from the pool, shaking droplets from their swim cap, and reaches out a hand to the fox, which responds with a playful nuzzle, highlighting an unusual yet heartwarming bond between human and wild animal. Throughout the scene, the lush greenery surrounding the pool provides a peaceful backdrop, contrasting with the vivid red of the swim cap and the rich orange tones of the fox's fur.]"
},
{
"from": "gpt",
"value": "Good. No obvious problems."
}
] |
seed44_2720 | videos/HRA_part1/2720.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A distinguished male doctor, wearing a white lab coat over casual attire, skillfully prepares dinner in a modern kitchen with stainless steel appliances and granite countertops. He expertly chops vegetables on a wooden cutting board, using a sharp chef's knife, while a pot simmers on the stove, releasing aromatic steam. The kitchen is illuminated by soft overhead lights, casting a warm glow over the scene. Behind him, a large window displays an urban skyline at twilight, adding a touch of sophistication to the domestic setting.]"
},
{
"from": "gpt",
"value": "Good. No obvious issues."
}
] |
seed44_1281 | videos/HRA_part1/1281.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A middle-aged janitor, wearing a gray jacket over his uniform, stands at the edge of a dense forest, contemplating his next move. He is holding hiking gear, including a walking stick, a backpack, and a water bottle, indicating his readiness for an outdoor adventure. The forest features tall, lush trees and a clear, sunny day, with a trail leading into the woods. The janitor begins his hike, walking confidently down the trail, his equipment in tow, as he explores the natural beauty of the forest surroundings.]"
},
{
"from": "gpt",
"value": "Normal. The character's face, hands, and legs are distorted in the video."
}
] |
seed46_1438 | videos/HRA_part2/1438.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A middle-aged man, donning a gray beanie, a dark jacket, and a checkered apron, stands against a vibrant blue wall, holding a large ladle as he stirs a bubbling pot. Moments later, he is seen attentively pouring oil into a pan, demonstrating culinary expertise. Then, he skillfully flips a food item in the pan, showcasing his cooking prowess, all while the blue wall and his attire remain consistent. Lastly, he is shown carefully seasoning the dish, his focus unwavering, against the same backdrop of the vivid blue wall and clad in the same attire, exemplifying his dedication to culinary artistry.]"
},
{
"from": "gpt",
"value": "Normal. Hand appears deformed."
}
] |
seed42_1593 | videos/HRA_part0/1593.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the semantic consistency issues in this video, focusing on the alignment between the text and the visual content. Specifically, point out any discrepancies regarding the subject (e.g., person or animal), quantity, color, scene description, style, and other relevant aspects. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A man in a gray shirt and blue jeans stands on a green ladder, cleaning out debris from the clogged gutters of a two-story house with a red roof, set against a clear blue sky. His focus is intense as he uses a small garden trowel to scoop out leaves and dirt. The house facade, painted in a warm beige hue, contrasts with the lush greenery surrounding it. A pile of removed debris sits on the ground beneath him, indicating progress in his task. His gray shirt and blue jeans are slightly soiled from the work, highlighting the effort and dedication put into maintaining the home's exterior.]"
},
{
"from": "gpt",
"value": "Normal. No debris was generated under the green ladder."
}
] |
seed42_1917 | videos/HRA_part0/1917.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the fidelity issues in this video, assessing the realism of the content, including people, animals, and other objects. Highlight any problems such as missing limbs, deformed hands or faces, or other unrealistic elements. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A beautifully cooked piece of salmon, expertly seared to achieve a perfect golden crust, sits center stage on a white ceramic plate. The salmon boasts a vibrant orange-pink interior, flaky yet moist, with a hint of dill sprinkled over it. Surrounding the fish are vibrant green asparagus spears, steamed to retain their crisp texture and natural color. Completing the dish is a drizzle of creamy sauce, likely a blend of lemon, butter, and capers, adding a touch of elegance and flavor. The plate is garnished with a sprig of parsley, enhancing the visual appeal of this exquisite salmon dish.]"
},
{
"from": "gpt",
"value": "Good. Basically correct."
}
] |
seed42_300 | videos/HRA_part0/300.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A majestic brown horse gallops across a lush, green meadow under a clear blue sky, its muscular body rippling with each powerful stride. The horse's mane and tail flow dramatically in the wind, creating a sense of speed and freedom. A pair of vibrant yellow flowers appear briefly in the foreground, adding a pop of color to the scene. The horse continues its run, hooves kicking up small clumps of earth, highlighting its strength and grace as it moves through the picturesque landscape.]"
},
{
"from": "gpt",
"value": "Normal. The video clearly shows the galloping horse, with basically normal movements, but there is slight disharmony."
}
] |
seed44_25 | videos/HRA_part1/25.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [In a whimsical scene, a friendly capybara, adorned with a miniature nurse's hat and a stethoscope around its neck, crawls across a colorful, childlike set filled with oversized medical-themed toys. The capybara moves deliberately, as if performing a check-up on a giant teddy bear patient lying on a playful examination table, surrounded by an array of oversized bandages, cotton balls, and a smiling face thermometer, creating a heartwarming and imaginative playtime scenario.]"
},
{
"from": "gpt",
"value": "Bad. The capybara's crawling movements in the video are not very continuous and smooth, showing some inconsistency."
}
] |
seed42_2468 | videos/HRA_part0/2468.mp4 | [
{
"from": "human",
"value": "<video>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, please watch the frames of a given video and see the text prompt for generating the video.\nThen identify the motion issues in this video, focusing on the continuity and smoothness of actions, as well as their coherence and adherence to the laws of physics. Provide a rating evaluation: Good, Normal, or Bad, along with a concise explanation supporting your assessment.\nYour task is provided as follows: Text Prompt: [A rugged worker, clad in a futuristic silver spacesuit with illuminated details, traverses the rocky, mineral-rich terrain of an asteroid under the faint light of distant stars. Carrying a high-tech pickaxe over his shoulder, he pauses to survey the alien landscape, showcasing determination and resilience in the harsh environment. As he walks, the textured surface of the asteroid and the tools strapped to his suit are clearly visible, emphasizing the challenging conditions of space mining. In the background, the vastness of space and the silhouette of a mining facility or spacecraft provide context to his solitary mission.]"
},
{
"from": "gpt",
"value": "Good. Basically no issues."
}
] |
End of preview. Expand
in Data Studio
LiFT-HRA
Dataset Summary
This dataset is derived from LiFT-HRA-20K for our UnifiedReward-7B training.
For further details, please refer to the following resources:
- 📰 Paper: https://arxiv.org/pdf/2503.05236
- 🪐 Project Page: https://codegoat24.github.io/UnifiedReward/
- 🤗 Model Collections: https://huggingface.co/collections/CodeGoat24/unifiedreward-models-67c3008148c3a380d15ac63a
- 🤗 Dataset Collections: https://huggingface.co/collections/CodeGoat24/unifiedreward-training-data-67c300d4fd5eff00fa7f1ede
- 👋 Point of Contact: Yibin Wang
Citation
@article{UnifiedReward,
title={Unified Reward Model for Multimodal Understanding and Generation.},
author={Wang, Yibin and Zang, Yuhang, and Li, Hao and Jin, Cheng and Wang Jiaqi},
journal={arXiv preprint arXiv:2503.05236},
year={2025}
}
- Downloads last month
- 50
Size of downloaded dataset files:
29 MB
Size of the auto-converted Parquet files:
7.99 MB
Number of rows:
19,009
Models trained or fine-tuned on CodeGoat24/LiFT-HRA