AI Paper Atlas

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

DiscoX: Benchmarking Discourse-Level Translation in Expert Domains

Xiying ZHAO, Zhoufutu Wen, Zhixuan Chen, Jingzhe Ding, Jianpeng Jiao, Shuai Li, Xi Li, Danni Liang, Shengda Long, Qianqian Liu, Xianbo Wu, Hongwan Gao and 10 more

ICLR 2026 · Large language models › Agents, code and math · approach: Mathematical and structural reasoning benchmarks

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
enable thinking mode (ablation)Qwen-3-235B · DiscoXoverall score: 59.66 -> 49.97Hurt
enable thinking mode (ablation)Claude-4 · DiscoXoverall score: 54.04 -> 53.53No clear effect
remove error de-duplication from Metric-S (ablation)Metric-S · human consistency, DiscoXaverage consistency: 66.00% -> 70.30%Improved
use a multi-agent workflow instead of a single LLM judge (ablation)Metric-S · human consistency, DiscoXaverage consistency: 24.90% -> 70.30%Improved
add fluency and appropriateness dimensions to accuracy-only judging (ablation)Metric-S · human consistency, DiscoXaverage consistency: 62.50% -> 70.30%Improved
use 60/20/20 dimension weights instead of equal weights (ablation)Metric-S · human consistency, DiscoXaverage consistency: 48.20% -> 70.30%Improved
use Gemini-2.5-Pro as judge instead of DeepSeek-R1 or o3-high (ablation)Metric-S · human consistency, DiscoXaverage consistency: 57.8% / 58.2% -> 70.3%Improved
o3 acting as judge shows self-preference (ablation)Metric-S with o3-high judge · DiscoX samplesjudge ranking: human 51.70 vs o3 judge 77.54 for o3-highHurt
compare translation direction en->zh vs zh->enLLMs · DiscoXoverall score: DeepSeek-V3 gap 34.8 points; Doubao-1.6-T gap 7.2Hurt
repeat evaluation in three independent runs (ablation)DiscoX · DiscoXoverall score: GPT-5: 76.66, 77.02, 77.03No clear effect

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