README.md

metadata

language:
  - en
  - zh
license: apache-2.0
datasets:
  - itsliupeng/mmlu_recall
model-index:
  - name: llama2_70b_mmlu
    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: 65.61
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co./spaces/HuggingFaceH4/open_llm_leaderboard?query=itsliupeng/llama2_70b_mmlu
          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: 87.37
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co./spaces/HuggingFaceH4/open_llm_leaderboard?query=itsliupeng/llama2_70b_mmlu
          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: 71.89
            name: accuracy
        source:
          url: >-
            https://huggingface.co./spaces/HuggingFaceH4/open_llm_leaderboard?query=itsliupeng/llama2_70b_mmlu
          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: 49.15
        source:
          url: >-
            https://huggingface.co./spaces/HuggingFaceH4/open_llm_leaderboard?query=itsliupeng/llama2_70b_mmlu
          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: 82.4
            name: accuracy
        source:
          url: >-
            https://huggingface.co./spaces/HuggingFaceH4/open_llm_leaderboard?query=itsliupeng/llama2_70b_mmlu
          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: 52.99
            name: accuracy
        source:
          url: >-
            https://huggingface.co./spaces/HuggingFaceH4/open_llm_leaderboard?query=itsliupeng/llama2_70b_mmlu
          name: Open LLM Leaderboard

We are utilizing the mmlu_recall dataset to continuously train the Llama-2-70b-hf model, aiming to enhance performance on mmlu metrics, while ensuring that other metric performances remain unaffected.

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	68.24
AI2 Reasoning Challenge (25-Shot)	65.61
HellaSwag (10-Shot)	87.37
MMLU (5-Shot)	71.89
TruthfulQA (0-shot)	49.15
Winogrande (5-shot)	82.40
GSM8k (5-shot)	52.99