GitHub - Fake10086/LLaVA-SAE

Name		Name	Last commit message	Last commit date
Latest commit History 4 Commits
llava		llava
README.md		README.md
save_tokenizer.py		save_tokenizer.py
train_sae.py		train_sae.py
train_sae.sh		train_sae.sh

Repository files navigation

A small self-contained SAE trained on LLaVA-v1.5-7B, and can be seamlessly used to interprete it.

train_sae.py corresponding to the .py to train and evaluate SAE.

./llava/model/language_model/sae.py corresponding to the .py to SAE model structure.

Everything else remains the same as the original LLaVA.

About

No description, website, or topics provided.

Report repository

Releases

No releases published

Packages

No packages published

Languages