Datasets:

---

CodeGoat24

/

LiFT-HRA

like 0

[ { "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 the heart of Shanghai, under the glow of neon signs and the hustle of the city streets, a peculiar scene unfolds. A street musician, dressed in a vibrant red jacket and dark jeans, passionately plays a saxophone, its golden finish shimmering under the streetlights. Beside him, a chimpanzee, wearing a tiny blue cap and a checkered shirt, mimics the musician's movements, adding an element of whimsy to the performance. As the music fills the air, a group of runners, clad in various sportswear and illuminated by the urban landscape, passes by, some pausing to watch the unique spectacle, while others continue their journey through the lively Shanghai night.]" }, { "from": "gpt", "value": "Bad. The musician's hand appears deformed, and people in the background also appear deformed." } ]

[ { "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 attractive male actor wearing a black t-shirt and dark jeans is seen running alongside a large, energetic yellow dog in an expansive park. The two are captured mid-stride, with the actor's focused gaze and the dog's joyful expression highlighting their shared activity. They run together in harmony, the actor's casual attire and the dog's bounding energy creating a dynamic and engaging scene against the lush greenery of the park.]" }, { "from": "gpt", "value": "Good. Basically correct." } ]

[ { "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 futuristic space opera performance, a farmer wearing a traditional straw hat and a patched apron over a checkered shirt runs across the stage under a backdrop of stars and planets. Simultaneously, a life-sized robotic tiger, designed with intricate circuitry patterns, races alongside him, both moving with a sense of urgency. The farmer's face is set in determination, and the tiger's robotic eyes glow with programmed intensity, highlighting the dramatic chase sequence unfolding against the cosmic setting.]" }, { "from": "gpt", "value": "Bad. The movements of the farmer and the tiger in the video are not smooth." } ]

[ { "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 vast, temperate shrubland during the golden hour, a capybara, the world's largest rodent, is seen running energetically across the landscape. The shrubland is dotted with various green shrubs and low-lying vegetation, providing a natural habitat for the capybara. The animal, with its brown fur and distinctive rodent features, bounds gracefully over the terrain, highlighting its agility and speed. As it moves, the capybara leaves a trail of displaced grass and dust, showcasing its vitality in this serene environment.]" }, { "from": "gpt", "value": "Good. No obvious issues." } ]

[ { "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: [Two young children, a boy and a girl, stand on a wooden chair in a cozy, lived-in room, their faces illuminated by the soft glow of a lamp. The boy, wearing a striped shirt and jeans, holds a new light bulb, while the girl, dressed in a pink sweater and leggings, assists him by holding a screwdriver. The boy carefully unscrews the old light bulb from the socket, while the girl looks on attentively. They work together, their teamwork and concentration evident in their focused expressions. The room around them is filled with toys and books, highlighting the everyday setting of this domestic task.]" }, { "from": "gpt", "value": "Bad. No screwdriver held by the girl or old bulb in the boy's hand was generated." } ]

[ { "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 rugged firefighter, clad in full gear and a yellow fire-resistant jacket, is kneeling by a blocked drain on a rainy day, amidst urban surroundings. He wields a long, coiled hose and a specialized tool, methodically working to unclog the drain. His focused expression and the rain-soaked environment underscore the challenging task at hand. Moments later, he pulls the clog from the drain, revealing a mix of debris, while water begins to flow freely again, thanks to his diligent efforts. Throughout, the firefighter's protective gear and the wet cityscape provide context to the scene.]" }, { "from": "gpt", "value": "Good. Basically correct." } ]

[ { "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 with a badge and a black belt, runs through a vast meadow filled with lush green grass and vibrant wildflowers. The landscape stretches to the horizon, dotted with distant trees. His expression is one of determination, eyes focused ahead, and he strides with purpose, leaving a faint trail in the dewy grass. The sun, partially obscured by a cloud, casts a soft, golden light over the scene, highlighting the officer's movement as he navigates the uneven terrain, demonstrating commitment and urgency in his duty.]" }, { "from": "gpt", "value": "Good. No obvious problems." } ]

[ { "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 the vast, sun-soaked scrublands, a juvenile giraffe with a sandy coat engages in an unusual behavior, moving slowly across the dry, cracked earth. Its spots blend into the arid landscape, and it crawls with a curious determination, perhaps exploring its environment or searching for food amidst the sparse vegetation. The scene is intimate, focusing solely on the giraffe's unique actions in the desolate yet beautiful wilderness.]" }, { "from": "gpt", "value": "Normal. The limbs of the giraffe in the video are missing, and there are issues with the facial generation." } ]

[ { "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, middle-aged farmer with a short beard, dressed in a green flannel shirt, brown vest, brown trousers, and brown boots, hikes across a lush, green pasture. The vibrant colors of his clothing contrast beautifully with the verdant landscape. In the first scene, he pauses to look up at the sky, which is partly cloudy, indicating a pleasant day. Then, he continues hiking, walking uphill towards a dense forest in the distance, the rolling hills providing a stunning backdrop. Next, he stops to observe a group of sheep grazing nearby, his presence conveying a sense of peace and harmony with nature. Lastly, he is seen walking along a narrow path bordered by tall grass, his figure becoming smaller as he moves further away, encapsulating the spirit of adventure and connection to the land.]" }, { "from": "gpt", "value": "Bad. The character's movements in the video are not smooth." } ]