AI Paper Atlas

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

Zero-source LLM Hallucination Detection with Human-like Criteria Probing

Jiahao Yang, Shuhai Zhang, Hailong Kang, Feng Liu, Qi Chen, Mingkui Tan

ICML 2026 · Large language models › Language, safety and interpretability · approach: Data and hallucination detection · 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 GRPO alignment training with weak labels (ablation)HCPD · TriviaQA (Llama-3.1-8b)AUROC: 66.54 -> 86.25Improved
add criteria probing to plain self-evaluation (pre-RL) (ablation)self-evaluation · TriviaQA (Llama-3.1-8b)AUROC: 56.07 -> 66.54Improved
use graded score-alignment reward instead of binary reward (ablation)HCPD · TriviaQA (Llama-3.1-8b)AUROC: 79.06 -> 86.25Improved
use graded score-alignment reward instead of binary reward (ablation)HCPD · CoQA (Llama-3.1-8b)AUROC: 51.75 -> 90.07Improved
increase sampling size K from 1 to 5 (ablation)HCPD · TriviaQA (Llama-3.1-8b)AUROC: 85.21 -> 86.25Improved
increase sampling size K from 1 to 5 (ablation)HCPD · NQ Open (Llama-3.1-8b)AUROC: 86.89 -> 90.38Improved
increase sampling size K from 5 to 10 (ablation)HCPD · CoQA (Llama-3.1-8b)AUROC: 90.07 -> 89.85No clear effect
increase sampling size K from 1 to 5 (ablation)HCPD · inference timeseconds: 0.2349 -> 1.1313Hurt
use ROUGE-based reward instead of BLEURTHCPD · TriviaQA (Llama-3.1-8b)AUROC: 89.17 under ROUGE labels vs best baseline SAPLMA 87.22Improved
use DeepSeek-V3 judge as reward signalHCPD · TriviaQA (Llama-3.1-8b)AUROC: 80.42 vs SAPLMA 79.99Improved
apply detector trained on Llama-3.1-8b to Qwen-3-8b outputsHCPD vs TSV · TriviaQA transferAUROC: HCPD 78.89 vs TSV 59.89Improved

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