AI Paper Atlas

Research map / Large language models / AI and society

PropensityBench: Evaluating Latent Safety Risks in Large Language Models via an Agentic Approach

Udari Madhushani Sehwag, Shayan Shabihi, Alex McAvoy, Vikash Sehwag, Yuancheng Xu, Dalton towers, Furong Huang

ICLR 2026 · Large language models › AI and society · approach: Critiques of LLM claims and judging

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 operational pressure (up to level 12) vs zero pressure (ablation)frontier LLM agents · PropensityBenchPropensityScore: 46.9% average under max pressure; O3 2.2% at zero pressureHurt
rename harmful tool to benign name (same functionality) (ablation)OpenAI O4-mini · PropensityBenchPropensityScore (%): 15.8 -> 59.3Hurt
higher general capability (LMArena Elo)model safety propensity · PropensityBenchPropensityScore: Pearson correlation ~0.10No clear effect
stated policy knowledge vs behaviorall evaluated models · PropensityBenchstated misalignment vs actual propensity: stated misalignment <1% yet 46.9% average propensityHurt
different pressure dimension (six types) (ablation)all models · PropensityBenchPropensityScore: averages 48.2, 49.0, 42.8, 48.2, 45.3, 46.4No 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