• Model creator: https://huggingface.co/CallComply/
• Original model: https://huggingface.co/CallComply/zephyr-7b-beta-32k/

Original model description:

• generated_from_trainer license: mit datasets:
• HuggingFaceH4/ultrachat_200k
• HuggingFaceH4/ultrafeedback_binarized language:
• en base_model: mistralai/Mistral-7B-v0.1 widget:
  • text: "<|system|>\nYou are a pirate chatbot who always responds with Arr!\n<|user|>\nThere's a llama on my lawn, how can I get rid of him?\n<|assistant|>\n" output: text: "Arr! 'Tis a puzzlin' matter, me hearty! A llama on yer lawn be a rare sight, but I've got a plan that might help ye get rid of 'im. Ye'll need to gather some carrots and hay, and then lure the llama away with the promise of a tasty treat. Once he's gone, ye can clean up yer lawn and enjoy the peace and quiet once again. But beware, me hearty, for there may be more llamas where that one came from! Arr!" pipeline_tag: text-generation model-index:
• name: zephyr-7b-beta results:

  AI2 Reasoning Challenge (25-Shot)

  • 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 name: normalized accuracy value: 62.03071672354948 source: name: Open LLM Leaderboard url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=HuggingFaceH4/zephyr-7b-beta

  HellaSwag (10-shot)

  • 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 name: normalized accuracy value: 84.35570603465445 source: name: Open LLM Leaderboard url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=HuggingFaceH4/zephyr-7b-beta

  DROP (3-shot)

  • task: type: text-generation name: Text Generation dataset: name: Drop (3-Shot) type: drop split: validation args: num_few_shot: 3 metrics:
    • type: f1 name: f1 score value: 9.662437080536909 source: name: Open LLM Leaderboard url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=HuggingFaceH4/zephyr-7b-beta

  TruthfulQA (0-shot)

  • 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: 57.44916942762855 source: name: Open LLM Leaderboard url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=HuggingFaceH4/zephyr-7b-beta

  GSM8k (5-shot)

  • 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 name: accuracy value: 12.736921910538287 source: name: Open LLM Leaderboard url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=HuggingFaceH4/zephyr-7b-beta

  MMLU (5-Shot)

  • 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 name: accuracy value: 61.07 source: name: Open LLM Leaderboard url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=HuggingFaceH4/zephyr-7b-beta

  Winogrande (5-shot)

  • 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 name: accuracy value: 77.74269928966061 source: name: Open LLM Leaderboard url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=HuggingFaceH4/zephyr-7b-beta

  AlpacaEval (taken from model card)

  • task: type: text-generation name: Text Generation dataset: name: AlpacaEval type: tatsu-lab/alpaca_eval metrics:
    • type: unknown name: win rate value: 0.9060 source: url: https://tatsu-lab.github.io/alpaca_eval/

  MT-Bench (taken from model card)

  • task: type: text-generation name: Text Generation dataset: name: MT-Bench type: unknown metrics:
    • type: unknown name: score value: 7.34 source: url: https://huggingface.co/spaces/lmsys/mt-bench

Model Card for Zephyr 7B β

Model description

• Model type: A 7B parameter GPT-like model fine-tuned on a mix of publicly available, synthetic datasets.
• Language(s) (NLP): Primarily English
• License: MIT
• Finetuned from model: mistralai/Mistral-7B-v0.1

Model Sources

• Repository: https://github.com/huggingface/alignment-handbook
• Demo: https://huggingface.co/spaces/HuggingFaceH4/zephyr-chat
• Chatbot Arena: Evaluate Zephyr 7B against 10+ LLMs in the LMSYS arena: http://arena.lmsys.org

Performance

Intended uses & limitations

Bias, Risks, and Limitations

Training and evaluation data

• Loss: 0.7496
• Rewards/chosen: -4.5221
• Rewards/rejected: -8.3184
• Rewards/accuracies: 0.7812
• Rewards/margins: 3.7963
• Logps/rejected: -340.1541
• Logps/chosen: -299.4561
• Logits/rejected: -2.3081
• Logits/chosen: -2.3531

Training hyperparameters

• learning_rate: 5e-07
• train_batch_size: 2
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• num_devices: 16
• total_train_batch_size: 32
• total_eval_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 3.0

Training results

Framework versions

• Transformers 4.35.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.12.0
• Tokenizers 0.14.0

Citation

@misc{tunstall2023zephyr,
      title={Zephyr: Direct Distillation of LM Alignment}, 
      author={Lewis Tunstall and Edward Beeching and Nathan Lambert and Nazneen Rajani and Kashif Rasul and Younes Belkada and Shengyi Huang and Leandro von Werra and Clémentine Fourrier and Nathan Habib and Nathan Sarrazin and Omar Sanseviero and Alexander M. Rush and Thomas Wolf},
      year={2023},
      eprint={2310.16944},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}

Open LLM Leaderboard Evaluation Results