metadata

license: apache-2.0
library_name: transformers
tags:
  - mergekit
  - merge
base_model:
  - NousResearch/DeepHermes-3-Llama-3-8B-Preview
  - allura-org/L3.1-8b-RP-Ink
model-index:
  - name: LThreePointOne-8B-HermesInk
    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: 40.31
            name: strict accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Triangle104/LThreePointOne-8B-HermesInk
          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: 31.48
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Triangle104/LThreePointOne-8B-HermesInk
          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: 17.22
            name: exact match
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Triangle104/LThreePointOne-8B-HermesInk
          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: 9.73
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Triangle104/LThreePointOne-8B-HermesInk
          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: 10.02
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Triangle104/LThreePointOne-8B-HermesInk
          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: 27.42
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Triangle104/LThreePointOne-8B-HermesInk
          name: Open LLM Leaderboard

Merge

This is a merge of pre-trained language models created using mergekit.

Merge Details

DeepHermes model merged with Ink RP model.

Credit to NousResearch for DeepHermes-3-Llama-3-8B-Preview and Allura-Org for L3.1-8b-RP-Ink.

Merge Method

This model was merged using the SLERP merge method.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

slices:
  - sources:
      - model: allura-org/L3.1-8b-RP-Ink
        layer_range: [0, 32]
      - model: NousResearch/DeepHermes-3-Llama-3-8B-Preview
        layer_range: [0, 32]
merge_method: slerp
base_model: NousResearch/DeepHermes-3-Llama-3-8B-Preview
parameters:
  t:
    - filter: self_attn
      value: [0, 0.5, 0.3, 0.7, 1]
    - filter: mlp
      value: [1, 0.5, 0.7, 0.3, 0]
    - value: 0.5
dtype: bfloat16

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	22.70
IFEval (0-Shot)	40.31
BBH (3-Shot)	31.48
MATH Lvl 5 (4-Shot)	17.22
GPQA (0-shot)	9.73
MuSR (0-shot)	10.02
MMLU-PRO (5-shot)	27.42