Safety & Alignment

Beyond Linear Probes: Dynamic Safety Monitoring for Language Models

J. Oldfield, P. Torr, I. Patras, A. Bibi, F. Barez

When AI Systems Learn During Deployment, Our Safety Evaluations Break

Best-of-N Jailbreaking

J. Hughes, S. Price, A. Lynch, R. Schaeffer, F. Barez, S. Koyejo, H. Sleight, E. Jones, E. Perez

Same Question, Different Words: A Latent Adversarial Framework for Prompt Robustness

T. Fu, F. Barez

Trust Me, I'm Wrong: High-Certainty Hallucinations in LLMs

A. Simhi, I. Itzhak, F. Barez, G. Stanovsky, Y. Belinkov

Chain-of-Thought Hijacking

J. Zhao, T. Fu, R. Schaeffer, M. Sharma, F. Barez

HACK: Hallucinations Along Certainty and Knowledge Axes

A. Simhi, J. Herzig, I. Itzhak, D. Arad, Z. Gekhman, R. Reichart, F. Barez, Et Al.

Rethinking Safety in LLM Fine-Tuning: An Optimization Perspective

M. Kim, J. M. Kwak, L. Alssum, B. Ghanem, P. Torr, D. Krueger, F. Barez†, A. Bibi†

Val-Bench: Measuring Value Alignment in Language Models

A. Gupta, D. O'Shea, F. Barez

PoisonBench: Assessing Large Language Model Vulnerability to Data Poisoning

T. Fu*, M. Sharma, P. Torr, S. B. Cohen, D. Krueger, F. Barez*