Azazelle/L3-RP_io

merge

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

Merge Details

Merge Method

This model was merged using the TIES merge method using meta-llama/Meta-Llama-3-8B-Instruct as a base.

Models Merged

The following models were included in the merge:

• ResplendentAI/Aura_Uncensored_l3_8B
• ResplendentAI/Kei_Llama3_8B
• Undi95/Llama-3-Unholy-8B
• vicgalle/Roleplay-Llama-3-8B

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: ResplendentAI/Aura_Uncensored_l3_8B
    parameters:
      density: 0.4
      weight: 0.4
  - model: ResplendentAI/Kei_Llama3_8B
    parameters:
      density: 0.4
      weight: 0.4
  - model: Undi95/Llama-3-Unholy-8B
    parameters:
      density: 0.3
      weight: 0.2
  - model: vicgalle/Roleplay-Llama-3-8B
    parameters:
      density: 0.3
      weight: 0.3
merge_method: ties
base_model: meta-llama/Meta-Llama-3-8B-Instruct
parameters:
  rescale: true
  normalize: false
  int8_mask: true
dtype: bfloat16

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	67.88
AI2 Reasoning Challenge (25-Shot)	63.05
HellaSwag (10-Shot)	79.86
MMLU (5-Shot)	67.92
TruthfulQA (0-shot)	52.90
Winogrande (5-shot)	75.69
GSM8k (5-shot)	67.85