AI Paper Atlas

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

HARP: Hallucination Detection via Reasoning Subspace Projection

Junjie Hu, Gang Tu, Cheng Shengyu, JinXin Li, Jinting Wang, Rui Chen, Zhilong Zhou, Dongbo Shan

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Task vectors and steering · 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
project hidden states onto reasoning subspace (vs raw features of same dimension) (ablation)HARP, Qwen-2.5-7B-Instruct · NQ OpenAUROC: 62.9 -> 84.0Improved
project hidden states onto reasoning subspace (vs raw features) (ablation)HARP, Qwen-2.5-7B-Instruct · TruthfulQAAUROC: 70.7 -> 88.1Improved
project hidden states onto reasoning subspace (vs raw features) (ablation)HARP, LLaMA-3.1-8B · NQ OpenAUROC: 70.4 -> 89.4Improved
use reasoning subspace basis instead of random basis vectors (ablation)HARP, Qwen-2.5-7B-Instruct · NQ OpenAUROC: 67.6 -> 84.0Improved
use reasoning subspace basis instead of random basis vectors (ablation)HARP, LLaMA-3.1-8B · TruthfulQAAUROC: 75.8 -> 88.5Improved
HARP vs best baselineHARP · TriviaQAAUROC: +7.5% (Qwen), +16.6% (LLaMA)Improved
set reasoning subspace dimension to 256 (about 5% of hidden size) (ablation)HARP · QA datasetsAUROC: range 32 to 1024 testedMixed
train detector on TriviaQA, test on NQ Open (ablation)HARP · NQ Open (cross-dataset)AUROCNo clear effect
intervene on reasoning components at deep layers (22, 23) (ablation)Qwen2.5 hidden states · Epsilon fictitious-capital demoAUROCImproved
intervene on reasoning components at shallow layers (1, 2) (ablation)Qwen2.5 hidden states · Epsilon fictitious-capital demoAUROCNo 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