AI Paper Atlas

Research map / Large language models / AI and society

LitReview Arena: Evaluating Literature Review Agents with Battle-style Peer Review Platform

Ruotong Zhao, zhiyu chen, Xurui Liu, Haidong Xue, Dong Liang, Jigao Fu, Wu YanBiao, Yuanyi Zhen, Fengli Xu, Yong Li

ICML 2026 · Large language models › AI and society · approach: Data commons and human baselines · 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
use agentic deep research systems instead of base language modelsliterature review generation · LitReviewBenchAverage overall utility Elo: 721.1 -> 1181.6Improved
use agentic systems (more test-time compute)literature review generation · LitReviewBenchtoken cost per query: 8.098K -> 122.297KHurt
allocate more computation to search and synthesisagentic systems · LitReviewBench vs human draftswin rate vs humans (D5): 23.0%No clear effect
increase draft length or interaction turnsliterature review agents · LitReviewBenchD3/D4 scoresNo clear effect
use an uncalibrated LLM judge on retrieval-like criterion (D1)Qwen3-235B judge · LitReviewBench D1Spearman rho with experts: 0.552Mixed
use an uncalibrated LLM judge on synthesis criteriaQwen3-235B judge · LitReviewBench D2-D4Spearman rho with experts: 0.442 (D2), 0.467 (D3), 0.430 (D4)Hurt
calibrate judge with structure-, content-similar demonstrations and gap anchors (LitJudge)Qwen3-235B judge · LitReviewBench D5alignment with experts: 0.467 -> 0.792Improved
add expert-derived gap anchors with MMR selection for D4LitJudge · LitReviewBench D4alignment with experts: 0.430 -> 0.842Improved
calibrate judge on Literature CoverageLitJudge · LitReviewBench D1alignment with experts: 0.552 -> 0.576Improved
calibrate judge for Claim Support and Paper StructureLitJudge · LitReviewBench D2, D3alignment with experts: 0.442 -> 0.673 (D2); 0.467 -> 0.649 (D3)Improved
use LitJudge calibration versus random few-shot ICL control (ablation)LitJudge · LitReviewBenchalignment with expertsImproved
use specialized systems Open Deep Research and SurveyForgeliterature review generation · LitReviewBenchexpert preferenceImproved

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