Search Results for author:
Found 2 papers, 1 papers with code
Break the Breakout: Reinventing LM Defense Against Jailbreak Attacks with Self-Refinement
We propose self-refine with formatting that achieves outstanding safety even in non-safety-aligned LMs and evaluate our method alongside several defense baselines, demonstrating that it is the safest training-free method against jailbreak attacks.
GTA: Gated Toxicity Avoidance for LM Performance Preservation
Our findings reveal that gated toxicity avoidance efficiently achieves comparable levels of toxicity reduction to the original CTG methods while preserving the generation performance of the language model.
Cannot find the paper you are looking for? You can
Submit a new open access paper.
