ColorShadow-7B

This is a Gradient-SLERP merge between diffnamehard/Mistral-CatMacaroni-slerp-7B and cookinai/Valkyrie-V1 performed using mergekit.

Here is the config file used:

  slices:
    - sources:
        - model: diffnamehard/Mistral-CatMacaroni-slerp-7B
          layer_range: [0, 32]
        - model: cookinai/Valkyrie-V1
          layer_range: [0, 32]
  merge_method: slerp
  base_model: diffnamehard/Mistral-CatMacaroni-slerp-7B
  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 # fallback for rest of tensors
  dtype: float16

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	68.34
AI2 Reasoning Challenge (25-Shot)	67.83
HellaSwag (10-Shot)	85.15
MMLU (5-Shot)	61.69
TruthfulQA (0-shot)	59.56
Winogrande (5-shot)	80.58
GSM8k (5-shot)	55.19