Jacoby746's picture
Adding Evaluation Results (#1)
57d83c5 verified
metadata
license: apache-2.0
base_model:
  - BAAI/Infinity-Instruct-7M-Gen-mistral-7B
  - Senseable/WestLake-7B-v2
  - SanjiWatsuki/Kunoichi-DPO-v2-7B
  - Gryphe/Tiamat-7b-1.1-DPO
  - uukuguy/speechless-instruct-mistral-7b-v0.2
base_model_relation: merge
model-index:
  - name: Proto-Athena-v0.2-4x7B
    results:
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: IFEval (0-Shot)
          type: HuggingFaceH4/ifeval
          args:
            num_few_shot: 0
        metrics:
          - type: inst_level_strict_acc and prompt_level_strict_acc
            value: 37.52
            name: strict accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Jacoby746/Proto-Athena-v0.2-4x7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: BBH (3-Shot)
          type: BBH
          args:
            num_few_shot: 3
        metrics:
          - type: acc_norm
            value: 30.34
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Jacoby746/Proto-Athena-v0.2-4x7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MATH Lvl 5 (4-Shot)
          type: hendrycks/competition_math
          args:
            num_few_shot: 4
        metrics:
          - type: exact_match
            value: 5.14
            name: exact match
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Jacoby746/Proto-Athena-v0.2-4x7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: GPQA (0-shot)
          type: Idavidrein/gpqa
          args:
            num_few_shot: 0
        metrics:
          - type: acc_norm
            value: 6.49
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Jacoby746/Proto-Athena-v0.2-4x7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MuSR (0-shot)
          type: TAUR-Lab/MuSR
          args:
            num_few_shot: 0
        metrics:
          - type: acc_norm
            value: 10.96
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Jacoby746/Proto-Athena-v0.2-4x7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MMLU-PRO (5-shot)
          type: TIGER-Lab/MMLU-Pro
          config: main
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 24.41
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Jacoby746/Proto-Athena-v0.2-4x7B
          name: Open LLM Leaderboard

Test merge of 7b models for learning purposees.

New in v0.2: Wanted to try a different gate type and using bfloat16, along with more detailed prompting to see if there's a noticeable difference.

Description: This model is a merge of BAAI/Infinity-Instruct-7M-Gen-mistral-7B, SanjiWatsuki/Kunoichi-7B, Gryphe_Tiamat-7b-1.1-DPO, Senseable_WestLake-7B-v2 and uukuguy/speechless-instruct-mistral-7b-v0.2 This is the first model I've ever uploaded and wanted to learn more about the process. Merged using mergekit-moe.

Works up to 8k context, 16k with 2.5 RoPe scaling

Prompt template: Custom format, or Alpaca

Alpaca: Below is an instruction that describes a task. Write a response that appropriately completes the request.

Instruction:

{prompt}

Response:

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric Value
Avg. 19.14
IFEval (0-Shot) 37.52
BBH (3-Shot) 30.34
MATH Lvl 5 (4-Shot) 5.14
GPQA (0-shot) 6.49
MuSR (0-shot) 10.96
MMLU-PRO (5-shot) 24.41