A Few Neurons Reveal When LLMs Misuse Tools: Sparse Detection and Selective Steering for Reliable Tool Use
Yutong Ke, Ming Yin, Chongwen Zhao, Kaizhu Huang
cs.CL
Submitted: 2026-07-31
Comments: 21 pages, 4 figures. Includes supplementary material
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Controlling Tool Use with Heading-Specific Activation Steering
- Model-Adaptive Tool Necessity Reveals the Knowing-Doing Gap in LLM Tool Use
- H-Neurons: On the Existence, Impact, and Origin of Hallucination-Associated Neurons in LLMs
- A Single Neuron Is Sufficient to Bypass Safety Alignment in Large Language Models
- TACT: Mitigating Overthinking and Overacting in Coding Agents via Activation Steering
- LLM Agents Already Know When to Call Tools -- Even Without Reasoning
- Steering Language Models With Activation Engineering
- ASA: Backbone-Training-Free Representation Engineering for Tool-Calling Agents
- Tool Calling is Linearly Readable and Steerable in Language Models
- 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