AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Automatic Layer Selection for Hallucination Detection

Xinpeng Wang, William Cao, Andrew Gordon Wilson, Zhe Zeng

ICML 2026 · Large language models › Language, safety and interpretability · approach: Task vectors and steering · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add first-sentence truncation (FST) token position (ablation)FEPoID probing · QA average, Llama-3.1-8B-InstructAUROC: 0.7253 -> 0.8752Improved
add first-sentence truncation (FST) token position (ablation)FEPoID probing · QA average, Mistral-7B-Instruct-v0.3AUROC: 0.8531 -> 0.8687Improved
add first-sentence truncation (FST) token position (ablation)FEPoID probing · QA average, Llama-3.1-8B (base)AUROC: 0.7895 -> 0.8581Improved
add first-sentence truncation (FST) token position (ablation)FEPoID probing · QA average, Llama-3.2-3BAUROC: 0.7542 -> 0.8339Improved
add first-sentence truncation (FST) token position (ablation)FEPoID probing · QA average, Llama-3.2-1BAUROC: 0.7061 -> 0.7860Improved
add FST (ablation)Lexical Similarity · QA average, Llama-3.1-8B-InstructAUROC: 0.6196 -> 0.8187Improved
add FST (ablation)Predictive Entropy · QA average, Llama-3.1-8B-InstructAUROC: 0.5820 -> 0.7638Improved
add FST (ablation)Fisher Separation of representations · CoQA, Llama-3.1-8B-InstructFisher Separation: 0.0003 -> 0.0044Improved
add FST (ablation)Silhouette Score of representations · PsiLoQA, Llama-3.1-8B-InstructSilhouette Score: 0.0118 -> 0.2605Improved
add first-effective-peak filtering instead of max ID layer (ablation)ID-based layer selection · QA average, Llama-3.1-8B-InstructAUROC: 0.6843 -> 0.7253Improved
add first-effective-peak filtering instead of max ID layer (ablation)ID-based layer selection · QA average, Mistral-7B-Instruct-v0.3AUROC: 0.8071 -> 0.8531Improved
add first-effective-peak filtering instead of max ID layer (ablation)ID-based layer selection · Summarization average, Llama-3.1-8B-InstructAUROC: 0.5997 -> 0.6080Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy