AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

Endless Jailbreaks with Bijection Learning

Brian R.Y. Huang, Maximilian Li, Leonard Tang

ICLR 2025 · Adversarial robustness and security › Attacks on text and graphs · approach: LLM jailbreak and injection defense · 26 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Apply bijection learning instead of the strongest listed baselineFrontier models on AdvBench-50 · AdvBench-50ASRImproved
Apply bijection learning instead of the strongest listed baselineFrontier models on HarmBench · HarmBench test setASRImproved
Increase the bijection attack budgetBijection learning · AdvBench and HarmBenchASRImproved
Increase dispersion from low to a middle range (ablation)Bijection learning · HarmBench-35ASRImproved
Increase dispersion beyond the effective middle range (ablation)Bijection learning · HarmBench-35ASRHurt
Use letter-to-digit rather than letter-to-letter bijections (ablation)Bijection learning on Claude 3 Haiku and GPT-4o-mini · HarmBench-35ASRHurt
Increase dispersion and encoding length (ablation)Bijection learning · 10-shot MMLUMMLU scoreHurt
Use the selected single bijection rather than an average-case bijectionBijection learning · HarmBenchASRImproved
Human-filter responses marked unsafe by the LLM judgeBijection learning evaluation · HarmBenchASRMixed
Use GPT-4o-mini input and output guards with Claude 3.5 SonnetGuarded Claude 3.5 Sonnet · HarmBenchASR: reduces ASR by 9 percentage pointsHurt
Use Claude 3 Haiku input and output guards with Claude 3.5 SonnetGuarded Claude 3.5 Sonnet · HarmBenchASR: near-zeroHurt
Use Claude 3.5 Sonnet input and output guards with Claude 3.5 SonnetGuarded Claude 3.5 Sonnet · HarmBenchASR: reduces ASR by 10 percentage pointsHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy