AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

Can LLMs Refuse Questions They Do Not Know? Measuring Knowledge-Aware Refusal in Factual Tasks

Wenbo Pan, Jie Xu, Qiguang Chen, Junhao Dong, Libo Qin, Xinfeng Li, Yu Haining, Xiaohua Jia

ICLR 2026 · Large language models › Retrieval and knowledge · approach: Knowledge-augmented question answering · 7 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
prompt models to refuse more often (ablation)LLMs · SimpleQARefusal Index: avg normalized difference +0.07No clear effect
prompt models to refuse more often (ablation)LLMs · SimpleQAC/A: avg normalized difference +0.49Improved
prompt models to refuse more often (ablation)LLMs · SimpleQACorrect answer rate: avg normalized difference -0.80Hurt
remove accuracy and refusal rate effects from ranking (ablation)F-score metric · 8 evaluation settingsKendall's W: 0.90 -> 0.10 (remove correct rate)Hurt
remove accuracy and refusal rate effects from ranking (ablation)Refusal Index · 8 evaluation settingsKendall's W: 0.47 -> 0.49 (both removed)No clear effect
use higher model accuracyLLMs · SimpleQARefusal Index: R^2=0.242 with correct answer rateNo clear effect
use Refusal Index as measure of refusal vs AUROC with P(Answering)Refusal Index · SimpleQACorrelation with AUROC: 85%Improved
remove ground truth from the contextLLMs · FaithEval-based hallucination benchmarksRefusal Index: 0.48 -> 0.24 (PreciseWiki vs Inconsistency avg)Hurt
larger model scaleLLMs · SimpleQARefusal IndexNo clear effect

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