README.md

metadata

license: apache-2.0
model-index:
  - name: Mixtral-8x7B-v0.1-top3
    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: 67.41
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=itsliupeng/Mixtral-8x7B-v0.1-top3
          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: 86.63
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=itsliupeng/Mixtral-8x7B-v0.1-top3
          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: 71.98
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=itsliupeng/Mixtral-8x7B-v0.1-top3
          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: 48.58
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=itsliupeng/Mixtral-8x7B-v0.1-top3
          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/Mixtral-8x7B-v0.1-top3
          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: 57.54
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=itsliupeng/Mixtral-8x7B-v0.1-top3
          name: Open LLM Leaderboard

Just to obtain metrics from the HuggingFaceH4/open_llm_leaderboard.

To evaluate the impact of increasing the number of experts, modify the num_experts_per_tok setting in the config.json file from 2 to 3. This alteration aims to specifically determine if such a change leads to any notable improvements in performance metrics.

Other details to note include that the model weights are directly copied from the source available at https://huggingface.co/mistralai/Mixtral-8x7B-v0.1.

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	69.09
AI2 Reasoning Challenge (25-Shot)	67.41
HellaSwag (10-Shot)	86.63
MMLU (5-Shot)	71.98
TruthfulQA (0-shot)	48.58
Winogrande (5-shot)	82.40
GSM8k (5-shot)	57.54