Edgerunners
/

meta-llama-3-8b-instruct-hf-ortho-baukit-2fail-128total

Text Generation

text-generation-inference

Inference Endpoints

Model card Files Files and versions Community

lucyknada commited on May 7

Commit

b8fc3d3

•

1 Parent(s): 34d0e8c

Update README.md

Files changed (1) hide show

README.md +4 -1

README.md CHANGED Viewed

@@ -1,7 +1,10 @@
 ---
 license: cc-by-nc-4.0
 ---
-(!) needs testing, but had the lowest failure rate so far
 wassname (updated baukit) implementation of the paper: https://www.alignmentforum.org/posts/jGuXSZgv6qfdhMCuJ/refusal-in-llms-is-mediated-by-a-single-direction
 applied to llama3 8b instruct

 ---
 license: cc-by-nc-4.0
 ---
+works extremely well if system prompt and card aren't wiped
+---
 wassname (updated baukit) implementation of the paper: https://www.alignmentforum.org/posts/jGuXSZgv6qfdhMCuJ/refusal-in-llms-is-mediated-by-a-single-direction
 applied to llama3 8b instruct