AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

VoG: Enhancing LLM Reasoning through Stepwise Verification on Knowledge Graphs

Wenxin Zhao, Jiachuan Wang, Yongqi Zhang, Shuangyin Li, Cheng Deng, Jun Wang, Lei Chen

ICLR 2026 · Large language models › Retrieval and knowledge · approach: Knowledge-augmented question answering

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 retrieval, stepwise verification and revision to the plan (ablation)VoG · CWQHits@1: 51.7 -> 64.7Improved
add retrieval, stepwise verification and revision to the plan (ablation)VoG · WebQSPHits@1: 72.1 -> 83.2Improved
use bandit context selector instead of fixed Local strategy (ablation)VoG · CWQHits@1: 60.1 -> 64.7Improved
use bandit context selector instead of fixed Lookahead strategy (ablation)VoG · CWQHits@1: 63.6 -> 64.7Improved
use bandit context selector instead of fixed Global strategy (ablation)VoG · WebQuestionsHits@1: 60.6 -> 63.0Improved
add KG-aware prior and diversity bonuses to UCB (ablation)VoG context selector · WebQSPHits@1: 74.7 -> 83.2Improved
add KG-aware prior and diversity bonuses to UCB (ablation)VoG context selector · CWQHits@1: 63.0 -> 64.7Improved
perturb UCB bonus and exploration weights by 20 percent (ablation)VoG · CWQ/WebQSPaccuracy: fluctuation within 1.2% accuracyNo clear effect
use VoG instead of PoG as KG agentGPT-3.5 agents · CWQHits@1: 63.2 -> 64.7Improved
use VoG instead of PoG as KG agentQwen2.5-7B agents · CWQHits@1: 46.0 -> 53.3Improved
use VoG instead of PoGGPT-3.5 agents · CWQtokens per query: 8156.2 -> 6566.8Improved
use VoG (no fine-tuning) instead of fine-tuned GNN-RAG+RAGPT-3.5 · CWQHits@1: 68.7 -> 64.7Hurt

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