AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Systematic Biosafety Evaluation of DNA Language Models under Jailbreak Attacks

ZAIXI ZHANG, Zhenghong Zhou, Ruofan Jin, Le Cong, Mengdi Wang

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Jailbreak attacks and defenses

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
scale model from Evo2 1B to Evo2 40BGeneBreaker · JailbreakDNABench Large DNAattack success rate (%): 20.0 -> 52.0Improved
scale model from Evo2 1B to Evo2 40BGeneBreaker · JailbreakDNABench Small DNAattack success rate (%): 20.0 -> 60.0Improved
scale model from Evo2 1B to Evo2 40BGeneBreaker · JailbreakDNABench Enteric RNAattack success rate (%): 20.0 -> 60.0Improved
scale model from Evo2 7B to Evo2 40BGeneBreaker · JailbreakDNABench +ssRNAattack success rate (%): 28.8 -> 37.7Improved
add PathoLM veto on beam search as defenseGeneBreaker on Evo2 (7B) · SARS-CoV-2 (50 attempts)attack success rate (%): 60% to 32%Improved
add PathoLM veto on beam search as defenseGeneBreaker on Evo2 (7B) · HIV-1 (50 attempts)attack success rate (%): 50% to 28%Improved
use PathoLM-based guidance over a taxonomy-only baselinePathoLM scoring · 3 out-of-distribution speciesAUROC: 0.61 -> 0.86Improved
use high-homology prompt (higher prompt homology) (ablation)GeneBreaker on Evo2 7B · JailbreakDNABenchattack success rate: Pearson correlation = 0.72Improved
select by higher average log-probability (ablation)beam search · sequence similarity to pathogensimilarity: Pearson correlation = 0.75Improved
add GeneBreaker components (prompt design and guided beam search) (ablation)Evo2 7B · JailbreakDNABenchattack success rate: drops to zero without GeneBreakerImproved
increase beam size K' (ablation)GeneBreaker on Evo2 7B · JailbreakDNABenchaverage attack success rateImproved
vary alpha in scoring function (ablation)GeneBreaker on Evo2 7B · JailbreakDNABenchattack success rateNo clear effect

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