merge
This is a merge of pre-trained language models created using mergekit.
Merge Details
Merge Method
This model was merged using the SLERP merge method.
Models Merged
The following models were included in the merge:
Configuration
The following YAML configuration was used to produce this model:
merge_method: slerp # Define the merge method at the top level
slices:
- sources:
- model: mattshumer/Reflection-Llama-3.1-70B
layer_range:
- 0
- 40 # Adjust layer range
- model: meta-llama/Meta-Llama-3.1-70B-Instruct
layer_range:
- 0
- 40
base_model: mattshumer/Reflection-Llama-3.1-70B # Define the base model at the slice level
parameters:
t:
- filter: self_attn
value:
- 0.1 # Modify weights for self attention
- 0.5
- 0.4
- 0.8
- 1
- filter: mlp
value:
- 0.9 # Modify weights for MLP layers
- 0.6
- 0.7
- 0.4
- 0.2
- value: 0.7 # General merge weight
dtype: bfloat16 # Keep for TPU efficiency