ForSureTesterSim's picture
Update README.md
f92a0ef verified
|
Raw History Blame Contribute Delete
417 Bytes
metadata
license: apache-2.0
base_model: meta-llama/Llama-3.1-8B
tags:
  - merge
  - mergekit
  - model_stock
  - aim
language:
  - en
pipeline_tag: text-generation

Llama-3.1-8B-AIM-ModelStock-v1

This model fuses meta-llama/Llama-3.1-8B-Instruct and deepseek-ai/DeepSeek-R1-Distill-Llama-8B using Model Stock, heavily stabilized by Activation-Informed Merging (AIM) ($\Omega = 0.4$) to protect foundational parameters.