merge
This is a merge of pre-trained language models created using mergekit.
Merge Details
Merge Method
This model was merged using the passthrough merge method.
Models Merged
The following models were included in the merge:
- /content/models/testing-method
Configuration
The following YAML configuration was used to produce this model:
slices:
- sources:
- model: "/content/models/testing-method"
layer_range: [0, 24]
- sources: # add middle layers with residuals scaled to zero
- model: "/content/models/testing-method"
layer_range: [8, 24]
parameters:
scale:
- filter: o_proj
value: 0.0
- filter: down_proj
value: 0.0
- value: 1.0
- sources:
- model: "/content/models/testing-method"
layer_range: [24, 32]
merge_method: passthrough
dtype: bfloat16
Open LLM Leaderboard Evaluation Results
Detailed results can be found here
Metric | Value |
---|---|
Avg. | 62.24 |
AI2 Reasoning Challenge (25-Shot) | 67.75 |
HellaSwag (10-Shot) | 81.09 |
MMLU (5-Shot) | 59.75 |
TruthfulQA (0-shot) | 60.41 |
Winogrande (5-shot) | 76.80 |
GSM8k (5-shot) | 27.67 |
- Downloads last month
- 2,697
Invalid base_model specified in model card
metadata. Needs to be a model id from
hf.co/models.
Evaluation results
- normalized accuracy on AI2 Reasoning Challenge (25-Shot)test set Open LLM Leaderboard67.750
- normalized accuracy on HellaSwag (10-Shot)validation set Open LLM Leaderboard81.090
- accuracy on MMLU (5-Shot)test set Open LLM Leaderboard59.750
- mc2 on TruthfulQA (0-shot)validation set Open LLM Leaderboard60.410
- accuracy on Winogrande (5-shot)validation set Open LLM Leaderboard76.800
- accuracy on GSM8k (5-shot)test set Open LLM Leaderboard27.670