highlighted merges
Collection
6 items
•
Updated
•
1
This is a merge of pre-trained language models created using mergekit.
This model was merged using the Model Stock merge method using Sao10K/L3-8B-Stheno-v3.2 as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
models:
- model: refuelai/Llama-3-Refueled
- model: cognitivecomputations/dolphin-2.9-llama3-8b
- model: cgato/L3-TheSpice-8b-v0.8.3
- model: NeverSleep/Llama-3-Lumimaid-8B-v0.1-OAS
- model: NousResearch/Hermes-2-Theta-Llama-3-8B
merge_method: model_stock
base_model: Sao10K/L3-8B-Stheno-v3.2
normalize: false
int8_mask: true
dtype: bfloat16
Detailed results can be found here
Metric | Value |
---|---|
Avg. | 26.78 |
IFEval (0-Shot) | 67.86 |
BBH (3-Shot) | 36.41 |
MATH Lvl 5 (4-Shot) | 9.21 |
GPQA (0-shot) | 7.38 |
MuSR (0-shot) | 10.97 |
MMLU-PRO (5-shot) | 28.87 |