conftrace_

jailbreaking

4 papers

Explore in graph

Co-occurring keywords · papers mentioning both

aligned llm (2) brittleness (1) low-rank modifications (1) loss signal (1) many-shot (1) negative demonstration (1) positive affirmation (1) decoding probabilities (1) misalignment rate (1) smaller models (1)

Papers

Adversarial Reasoning at Jailbreaking Time ICML 2025 PANDAS: Improving Many-shot Jailbreaking via Positive Affirmation, Negative Demonstration, and Adaptive Sampling ICML 2025 Weak-to-Strong Jailbreaking on Large Language Models ICML 2025 Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications ICML 2024