The Safety Relay in Roleplay Jailbreaks: A Component-Resolved Causal Analysis of Harm Recognition and Refusal
cs.LG
Submitted: 2026-08-31
Updated: 2026-08-31
Terminology
Sources
- The Llama 3 Herd of Models
- DeepInception: Hypnotize Large Language Model to Be Jailbreaker
- Gemma 2: Improving Open Language Models at a Practical Size
- Representation Engineering: A Top-Down Approach to AI Transparency
- Scalable and Transferable Black-Box Jailbreaks for Language Models via Persona Modulation
- Universal and Transferable Adversarial Attacks on Aligned Language Models
- Qwen2.5 Technical Report
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks