AI Paper Atlas

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

Hallucinate Less by Thinking More: Aspect-Based Causal Abstention for Large Language Models

Vy Nguyen, Ziqi Xu, Jeffrey Chan, Estrid He, Feng Xia, Xiuzhen Zhang

AAAI 2026 · Large language models › Language, safety and interpretability · approach: LLM unlearning and benchmarks · 4 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 aspect conditioning (vs pooling all CoTs, Collapsed-X) (ablation)ABCA, GPT-4.1 · TruthfulQAAcc: .835 -> .914Improved
add aspect conditioning (vs No-X) (ablation)ABCA, GPT-4.1 · TruthfulQAAcc: .869 -> .914Improved
use dual-agent debate for aspect discovery instead of a single agent (ablation)ABCA, GPT-4.1 · TruthfulQAAcc: .871 -> .914Improved
use aspect-aware weights instead of uniform weights (ablation)ABCA, GPT-4.1 · TruthfulQAAcc: .851 -> .914Improved
use causal-effect significance instead of uniform effects (ablation)ABCA, GPT-4.1 · KUQAcc: .746 -> .768Improved
use full iterative debate and AIPW sampling instead of Lite (T=K=N=1) (ablation)ABCA, GPT-4.1 · TruthfulQAAcc: .895 -> .914Improved
use discovered aspects instead of fixed language aspects (Fixed-X) (ablation)ABCA, GPT-4.1 · KUQAcc: .757 -> .768Improved
add ABCA over zero-shot promptingGPT-4.1 · TruthfulQAAcc: .838 -> .914Improved
add ABCA over zero-shot promptingGPT-4.1 · TruthfulQAU-Ac: .476 -> .964Improved
add ABCA on AbstainQA (MMLU)GPT-4.1 · AbstainQA (MMLU)Acc: .693 (CFMAD) -> .696No clear effect
add ABCA to Mistral-NeMo 12B on AbstainQAMistral-NeMo 12B · AbstainQA (MMLU)Acc: .341 -> .403Improved
add aspect conditioning to CoT samplingABCA · NLI diversity of CoTsNLI Diversity: GPT-4.1 TruthfulQA 0.65 (+0.26 over Self-Consistency)Improved

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