A Safe Prototype Is Not a Safety Direction: Reference Dependence and Prompt Confounds in Response-Safety Embeddings
cs.LG, cs.CL, cs.CR
Submitted: 2026-10-01
Updated: 2026-10-01
Terminology
Sources
- Efficient LLM Moderation with Multi-Layer Latent Prototypes
- Benchmarking and Improving Monitors for Out-Of-Distribution Alignment Failure in LLMs
- Aegis2.0: A Diverse AI Safety Dataset and Risks Taxonomy for Alignment of LLM Guardrails
- Rethinking Jailbreak Detection of Large Vision Language Models with Representational Contrastive Scoring
- Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations
- Safe-Embed: Unveiling the Safety-Critical Knowledge of Sentence Encoders
- Qwen2.5 Technical Report
- Detecting Sleeper Agents in Large Language Models via Semantic Drift Analysis
- ShieldGemma: Generative AI Content Moderation Based on Gemma
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