Read-Best Is Not Steer-Best: A Probing--Steering Layer Dissociation in Omni-Modal Large Language Models

arXiv:2609.22135 · cs.CL · Submitted 2026-08-24 · Read on arXiv

cs.CL

Submitted: 2026-08-24

Updated: 2026-08-24

Comments: 13 pages, 10 figures

Code: https://github.com/YiboWang2002/Read-Best-Is-Not-Steer-Best

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Steering — 6× in this paper · explained in 2 episodes
• Accuracy — 2× in this paper · explained in 1 episode
• Activation Steering — 1× in this paper · explained in 4 episodes

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