T145/ZEUS-8B-V17

ZEUS 8B 🌩️ V17

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

Merge Details

Merge Method

This model was merged using the DARE TIES merge method using unsloth/Meta-Llama-3.1-8B-Instruct as a base.

Models Merged

The following models were included in the merge:

• unsloth/Llama-3.1-Storm-8B
• Orenguteng/Llama-3.1-8B-Lexi-Uncensored-V2
• VAGOsolutions/Llama-3.1-SauerkrautLM-8b-Instruct
• arcee-ai/Llama-3.1-SuperNova-Lite

Configuration

The following YAML configuration was used to produce this model:

base_model: unsloth/Meta-Llama-3.1-8B-Instruct
dtype: bfloat16
merge_method: dare_ties
parameters:
  int8_mask: 1.0
  normalize: 1.0
  random_seed: 145.0
slices:
- sources:
  - layer_range: [0, 32]
    model: unsloth/Llama-3.1-Storm-8B
    parameters:
      density: 0.95
      weight: 0.28
  - layer_range: [0, 32]
    model: arcee-ai/Llama-3.1-SuperNova-Lite
    parameters:
      density: 0.9
      weight: 0.27
  - layer_range: [0, 32]
    model: VAGOsolutions/Llama-3.1-SauerkrautLM-8b-Instruct
    parameters:
      density: 0.92
      weight: 0.25
  - layer_range: [0, 32]
    model: Orenguteng/Llama-3.1-8B-Lexi-Uncensored-V2
    parameters:
      density: 0.92
      weight: 0.2
  - layer_range: [0, 32]
    model: unsloth/Meta-Llama-3.1-8B-Instruct
tokenizer_source: union

Open LLM Leaderboard Evaluation Results

Detailed results can be found here! Summarized results can be found here!

Metric	Value (%)
Average	30.79
IFEval (0-Shot)	79.41
BBH (3-Shot)	32.34
MATH Lvl 5 (4-Shot)	21.15
GPQA (0-shot)	9.62
MuSR (0-shot)	9.64
MMLU-PRO (5-shot)	32.61