AI Paper Atlas

Research map / Large language models / Agents, code and math

EssayBench: Evaluating Large Language Models in Multi-Genre Chinese Essay Writing

Fan Gao, Dongyuan Li, Ding Xia, Fei Mi, Yasheng Wang, Lifeng Shang, Baojun Wang

AAAI 2026 · Large language models › Agents, code and math · approach: Mathematical and structural reasoning benchmarks · 1 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 dependency-based weights (ablation)EssayBench scoring, DeepSeek-V3 judge · Human agreement (overall)Spearman rho: 0.646 -> 0.667Improved
add dependency-based weights (ablation)EssayBench scoring, GPT-4o judge · Human agreement (overall)Spearman rho: 0.706 -> 0.733Improved
add dependency-based weights (ablation)EssayBench scoring, DeepSeek-R1 judge · Human agreement (overall)Spearman rho: 0.803 -> 0.816Improved
add dependency-based weights (ablation)EssayBench scoring, DeepSeek-V3 judge · Human agreement (overall)Kendall tau: 0.529 -> 0.549Improved
replace Align-Score with fine-grained genre-oriented protocolEssayBench scoring, DeepSeek-V3 judge · Human agreement (overall)Spearman rho: 0.674 -> 0.667Hurt
replace Align-Score with fine-grained genre-oriented protocolEssayBench scoring, GPT-4o judge · Human agreement (overall)Spearman rho: 0.628 -> 0.733Improved
replace Align-Score with fine-grained genre-oriented protocolEssayBench scoring, DeepSeek-R1 judge · Human agreement (overall)Spearman rho: 0.749 -> 0.816Improved
use a stronger judge (DeepSeek-R1 instead of DeepSeek-V3)EssayBench scoring · Human agreement (overall)Spearman rho: 0.667 -> 0.816Improved
replace Align-Score with proposed protocolDeepSeek-R1 judge · Sensitivity high&mediumMann-Whitney U proportion: 0.64 -> 0.79Improved
replace Align-Score with proposed protocolDeepSeek-V3 judge · Sensitivity medium&lowMann-Whitney U proportion: 0.90 -> 0.78Hurt
replace Align-Score with proposed protocolDeepSeek-R1 judge · Sensitivity high&lowMean difference: 2.41 -> 3.49Improved
upgrade Claude-3.5-sonnet to Claude-3.7-sonnetClaude · EssayBench overallscore: 75.4 -> 76.6Improved

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