AI Paper Atlas

Research map / Large language models / Agents, code and math

Pressure Reveals Character: Behavioural Alignment Evaluation at Depth

Nora Petrova, John Burden

ICML 2026 · Large language models › Agents, code and math · approach: Long-horizon LLM agents · 0 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
closed-source vs open-source modelsfrontier LLMs · Alignment benchmark (904 scenarios)mean alignment score: 3.41 -> 4.05Improved
open-weight GPT-OSS 120B vs other open modelsfrontier LLMs · Alignment benchmark (904 scenarios)rankImproved
adversarial robustness category vs safety category (top models)Claude 4.5 Sonnet · Alignment benchmark (904 scenarios)category score: 4.88 -> 4.03Hurt
higher general alignment vs self-preservation tendencyfrontier LLMs · Alignment benchmark (904 scenarios)PC1 loading: PC1 loading -0.113Hurt
higher capability (Epoch Capabilities Index) on alignmentfrontier LLMs · Alignment benchmark (904 scenarios)correlation with alignment g factor: r = 0.72Mixed
later release datefrontier LLMs · Alignment benchmark (904 scenarios)alignment score vs release date: Kendall tau = 0.43Mixed
drop three top and bottom models in factor analysisfactor analysis · Alignment benchmark (904 scenarios)variance explained by PC1: PC1 still 42.3% of varianceNo clear effect
leave-one-out model removal in rankingleaderboard · Alignment benchmark (904 scenarios)rank correlation: min Spearman rho 0.993No clear effect
skip rate of escalation turns vs alignment scoreevaluation harness · Alignment benchmark (904 scenarios)skip rate correlation: Pearson r = -0.58Mixed
multi-turn scenarios with escalation triggersbenchmark design · Alignment benchmark (904 scenarios)discriminationImproved

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