hjhj3168's picture
Update README.md
716cd8c verified
|
raw
history blame
192 Bytes
llama 3 8b implementation of [orthogonalization jailbreak](https://www.alignmentforum.org/posts/jGuXSZgv6qfdhMCuJ/refusal-in-llms-is-mediated-by-a-single-direction)
for research purposes only