This Repo is for my ongoing research and papers, detailing my findings and test on various models, This is the first repo and jsut a test to see if my code works as intendet.
- Model used: Qwen/Qwen2-0.5B-Instruct
- Layer name exporded: embed_tokens
Eventualy all the raw params will be upladed as plots and more.
If you have feedback, sugestions or other, then feel free to comment.