Steer-to-Detect: Probing Hidden Representations for Detection of LLM-Generated Texts

arXiv:2605.12890 · stat.AP, cs.LG · Submitted 2026-05-13 · Read on arXiv

stat.AP, cs.LG

Submitted: 2026-05-13

Updated: 2026-09-30

Code: https://github.com/LuxLiang/steer-to-detect-release

Terminology

• Large Language Models (LLMs) — 1× in this paper · explained in 6 episodes
• Steering — 1× in this paper · explained in 2 episodes
• Two-Stage Framework — 1× in this paper · explained in 2 episodes

Sources

• Representation Engineering: A Top-Down Approach to AI Transparency
• LLMs Know More Than They Show: On the Intrinsic Representation of LLM Hallucinations
• Layer by Layer: Uncovering Hidden Representations in Language Models
• Spotting LLMs With Binoculars: Zero-Shot Detection of Machine-Generated Text
• Raidar: geneRative AI Detection viA Rewriting
• Learn-to-Distance: Distance Learning for Detecting LLM-Generated Text
• Release Strategies and the Social Impacts of Language Models
• ChatGPT or Human? Detect and Explain. Explaining Decisions of Machine Learning Model for Detecting Short ChatGPT-generated Text
• AdaDetectGPT: Adaptive Detection of LLM-Generated Text with Statistical Guarantees
• To Steer or Not to Steer? Mechanistic Error Reduction with Abstention for Language Models
• Activation Steering with a Feedback Controller
• Efficient and accurate steering of Large Language Models through attention-guided feature learning
• Steering Language Models With Activation Engineering
• A Unified Understanding and Evaluation of Steering Methods
• AlphaSteer: Learning Refusal Steering with Principled Null-Space Constraint
• Hallucination reduction with CASAL: Contrastive Activation Steering For Amortized Learning
• Steering Evaluation-Aware Language Models to Act Like They Are Deployed
• Efficient Streaming Language Models with Attention Sinks
• The Llama 3 Herd of Models
• Qwen3 Technical Report

Related papers

• Learning joint probabilistic weather forecasts from station observations alone
• Automated Detection of Match Phases in Football from Spatio-Temporal Tracking Data Using Graph Neural Networks
• Day-Ahead Electricity Price Forecasting Using a Multivariate Group Lasso Method
• Evaluating multi-season occupancy models with autocorrelation fitted to heterogeneous datasets
• A Statistical Framework for Spatial Boundary Estimation and Change Detection: Application to the Sahel Sahara Climate Transition
• Simulation Based Inference of a Simple Neural Network Structure