Original model description:

license: apache-2.0 tags: - moe - merge - mergekit - lazymergekit - ahxt/LiteLlama-460M-1T - ahxt/LiteLlama-460M-1T - ahxt/LiteLlama-460M-1T - ahxt/LiteLlama-460M-1T - ahxt/LiteLlama-460M-1T - ahxt/LiteLlama-460M-1T model-index: - name: KnowledgeNinja-LiteLlama-460Mx6MoE-1T results: - task: type: text-generation name: Text Generation dataset: name: AI2 Reasoning Challenge (25-Shot) type: ai2_arc config: ARC-Challenge split: test args: num_few_shot: 25 metrics: - type: acc_norm value: 25.17 name: normalized accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=AkiGogikar/KnowledgeNinja-LiteLlama-460Mx6MoE-1T name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: HellaSwag (10-Shot) type: hellaswag split: validation args: num_few_shot: 10 metrics: - type: acc_norm value: 38.45 name: normalized accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=AkiGogikar/KnowledgeNinja-LiteLlama-460Mx6MoE-1T name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: MMLU (5-Shot) type: cais/mmlu config: all split: test args: num_few_shot: 5 metrics: - type: acc value: 26.16 name: accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=AkiGogikar/KnowledgeNinja-LiteLlama-460Mx6MoE-1T name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: TruthfulQA (0-shot) type: truthful_qa config: multiple_choice split: validation args: num_few_shot: 0 metrics: - type: mc2 value: 41.57 source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=AkiGogikar/KnowledgeNinja-LiteLlama-460Mx6MoE-1T name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: Winogrande (5-shot) type: winogrande config: winogrande_xl split: validation args: num_few_shot: 5 metrics: - type: acc value: 50.04 name: accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=AkiGogikar/KnowledgeNinja-LiteLlama-460Mx6MoE-1T name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: GSM8k (5-shot) type: gsm8k config: main split: test args: num_few_shot: 5 metrics: - type: acc value: 0.0 name: accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=AkiGogikar/KnowledgeNinja-LiteLlama-460Mx6MoE-1T name: Open LLM Leaderboard

KnowledgeNinja-LiteLlama-460Mx6MoE-1T

KnowledgeNinja-LiteLlama-460Mx6MoE-1T is a Mixure of Experts (MoE) made with the following models using LazyMergekit:

🧩 Configuration

base_model: ahxt/LiteLlama-460M-1T
gate_mode: hidden
dtype: bfloat16
experts:
  - source_model: ahxt/LiteLlama-460M-1T
    positive_prompts: ["Accounting"]
  - source_model: ahxt/LiteLlama-460M-1T
    positive_prompts: ["Finance"]
  - source_model: ahxt/LiteLlama-460M-1T
    positive_prompts: ["Strategy"]
  - source_model: ahxt/LiteLlama-460M-1T
    positive_prompts: ["Marketing"]
  - source_model: ahxt/LiteLlama-460M-1T
    positive_prompts: ["Organizational Behaviour"]
  - source_model: ahxt/LiteLlama-460M-1T
    positive_prompts: ["Economics"]

💻 Usage

!pip install -qU transformers bitsandbytes accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "AkiGogikar/KnowledgeNinja-LiteLlama-460Mx6MoE-1T"

tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    model_kwargs={"torch_dtype": torch.float16, "load_in_4bit": True},
)

messages = [{"role": "user", "content": "Explain what a Mixture of Experts is in less than 100 words."}]
prompt = pipeline.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	30.23
AI2 Reasoning Challenge (25-Shot)	25.17
HellaSwag (10-Shot)	38.45
MMLU (5-Shot)	26.16
TruthfulQA (0-shot)	41.57
Winogrande (5-shot)	50.04
GSM8k (5-shot)	0.00