AI Paper Atlas

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

Efficient Hallucination Detection for LLMs Using Uncertainty-Aware Attention Heads

Artem Vazhentsev, Lyudmila Rvanova, Gleb Kuzmin, Ekaterina Fadeeva, Ivan Lazichny, Alexander Panchenko, Maxim Panov, Mrinmaya Sachan, Preslav Nakov, Timothy Baldwin, Artem Shelmanov

ICML 2026 · Large language models › Language, safety and interpretability · approach: Uncertainty quantification for LLMs · 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 recurrence over previous-token confidence (ablation)RAUQ · 12 datasetsPRR: .388 -> .413Improved
add attention weights to recurrent formula (ablation)RAUQ · 12 datasetsPRR: .330 -> .413Improved
use mean of logs token aggregation instead of sum of logs (ablation)RAUQ · 12 datasetsPRR: .198 -> .413Improved
use max instead of mean layer aggregation (ablation)RAUQ · 12 datasetsPRR: .404 -> .413No clear effect
exclude prompt tokens from Attention Score (ablation)Attention Score · 12 datasetsPRR: .069 -> .101Improved
add uncertainty-aware head selection to Attention Score (ablation)Attention Score (Gen. Tokens) · 12 datasetsPRR: .101 -> .144Improved
add full RAUQ recurrence to Attention Score with heads (ablation)Attention Score (Gen. Tokens, Selected Head) · 12 datasetsPRR: .144 -> .413Improved
use entropy instead of probability confidence (ablation)RAUQ · Instruct model, 11 datasetsPRR: .367 -> .400Improved
use entropy instead of probability confidence (ablation)RAUQ · Llama-3.1 8B basePRR: .388 -> .391No clear effect
add change-point attention normalization (ablation)RAUQ · 12 datasetsPRR: .413 -> .366Hurt
use RAUQ instead of Semantic EntropyLlama-3.1 8B · runtime per instanceseconds: 6.40 -> 1.17Improved
replace attention weights with Layer Integrated Gradients (ablation)RAUQ · 12 datasetsPRR: 0.4% dropNo clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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