Over-Refusal and Representation Subspaces: A Mechanistic Analysis of Task-Conditioned Refusal in Aligned LLMs

arXiv:2603.27518 · cs.CL · Submitted 2026-03-29 · Read on arXiv

cs.CL

Submitted: 2026-03-29

Updated: 2026-09-01

Terminology

• Steering — 32× in this paper · explained in 2 episodes
• Ablation — 29× in this paper · explained in 1 episode
• Accuracy — 5× in this paper · explained in 1 episode
• Linear Probing — 5× in this paper · explained in 2 episodes
• RAG — 5× in this paper · explained in 3 episodes
• Transformer — 5× in this paper · explained in 2 episodes
• Benchmark — 4× in this paper · explained in 2 episodes
• Interpretability — 4× in this paper · explained in 15 episodes
• Activation Steering — 3× in this paper · explained in 4 episodes
• Alignment — 3× in this paper · explained in 1 episode
• Cosine Similarity — 3× in this paper · explained in 3 episodes
• Fine-Tuning — 3× in this paper · explained in 5 episodes
• Adam — 2× in this paper · explained in 2 episodes
• Embeddings — 2× in this paper · explained in 2 episodes
• Linear probe — 2× in this paper · explained in 2 episodes
• Representation Engineering — 2× in this paper · explained in 2 episodes
• Abstention — 1× in this paper · explained in 1 episode
• Activation Addition — 1× in this paper · explained in 1 episode
• Activation Patching — 1× in this paper · explained in 2 episodes
• Core — 1× in this paper · explained in 2 episodes
• Decode — 1× in this paper · explained in 2 episodes
• Foundation Model — 1× in this paper · explained in 3 episodes
• Foundation Models — 1× in this paper · explained in 7 episodes
• Grounding — 1× in this paper · explained in 3 episodes
• Hallucination — 1× in this paper · explained in 3 episodes
• Over Refusal — 1× in this paper
• Robustness — 1× in this paper · explained in 5 episodes
• Structured Reasoning — 1× in this paper · explained in 3 episodes
• Trace — 1× in this paper · explained in 2 episodes

Sources

• SafeSwitch: Steering Unsafe LLM Behavior via Internal Activation Signals
• Uncovering Intermediate Variables in Transformers using Circuit Probing
• SafeConstellations: Mitigating Over-Refusals in LLMs Through Task-Aware Representation Steering
• The Linear Representation Hypothesis and the Geometry of Large Language Models
• Refusal Falls off a Cliff: How Safety Alignment Fails in Reasoning?
• Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models
• Understanding and Mitigating Over-refusal for Large Language Models via Safety Representation
• Representation Engineering: A Top-Down Approach to AI Transparency

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering