From Expected Harmfulness to Likelihood: A Probabilistic Reformulation of Jailbreaking LLM Agents
cs.AI
Submitted: 2026-10-07
Updated: 2026-10-07
Code: https://github.com/Hax1on/OPUR
Sources
- Expected Harm: Rethinking Safety Evaluation of (Mis)Aligned LLMs
- MasterKey: Automated Jailbreak Across Multiple Large Language Model Chatbots
- REINFORCE Adversarial Attacks on Large Language Models: An Adaptive, Distributional, and Semantic Objective
- The Llama 3 Herd of Models
- COLD-Attack: Jailbreaking LLMs with Stealthiness and Controllability
- Mistral 7B
- DeepInception: Hypnotize Large Language Model to Be Jailbreaker
- AmpleGCG: Learning a Universal and Transferable Generative Model of Adversarial Suffixes for Jailbreaking Both Open and Closed LLMs
- A Trembling House of Cards? Mapping Adversarial Attacks against Language Agents
- AdvPrompter: Fast Adaptive Adversarial Prompting for LLMs
- Fast Adversarial Attacks on Language Models In One GPU Minute
- ReAct: Synergizing Reasoning and Acting in Language Models
- UDora: A Unified Red Teaming Framework against LLM Agents by Dynamically Hijacking Their Own Reasoning
- Universal and Transferable Adversarial Attacks on Aligned Language Models
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems