AI Paper Atlas

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

Can LLMs Detect Their Confabulations? Estimating Reliability in Uncertainty-Aware Language Models

Tianyi Zhou, Johanne Medina, Sanjay Chawla

AAAI 2026 · Large language models › Language, safety and interpretability · approach: Uncertainty quantification for LLMs · 2 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
replace LogProb with uncertainty-guided Probe(AVG)Fanar1-9b · TruthfulQAAUROC: 0.597 -> 0.734Improved
replace LogProb with uncertainty-guided Probe(AVG)Fanar1-9b · MathAUROC: 0.757 -> 0.833Improved
replace LogProb with Probe(AVG)Qwen2.5-7B · TruthfulQAAUROC: 0.591 -> 0.761Improved
replace LogProb with Probe(AVG)Qwen2.5-7B · MathAUROC: 0.635 -> 0.699Improved
replace LogProb with Probe(AVG)Gemma3-12B · TruthfulQAAUROC: 0.545 -> 0.733Improved
replace LogProb with Probe(AVG)Gemma3-12B · TriviaQAAUROC: 0.806 -> 0.818Improved
replace LogProb with Probe(AVG)Fanar1-9b · TriviaQAAUROC: 0.774 -> 0.765Hurt
replace LogTokU with Probe(AVG)Gemma3-12B · MathAUROC: 0.791 -> 0.786Hurt
average hidden states of selected tokens (Probe(AVG)) instead of EOS token only (ablation)Fanar1-9b · TruthfulQAAUROC: 0.706 -> 0.734Improved
average hidden states (Probe(AVG)) instead of EOS token only (ablation)Qwen2.5-7B · TriviaQAAUROC: 0.812 -> 0.786Hurt
average hidden states (Probe(AVG)) instead of single highest/lowest EU token (Probe(EU)) (ablation)Gemma3-12B · MathAUROC: 0.669 -> 0.786Improved
average hidden states (Probe(AVG)) instead of single EU token (Probe(EU)) (ablation)Fanar1-9b · TriviaQAAUROC: 0.751 -> 0.765Improved

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