AI Paper Atlas

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

Compress-then-Rank: Faster and Better Listwise Reranking with Large Language Models via Ranking-Aware Passage Compression

Zhewei Zhi, Yingyi Zhang, Yizhen Jing, Xianneng Li, Jianing Liu, Huajie Liu, Yongliang Ding

AAAI 2026 · Large language models › Language, safety and interpretability · approach: LLM pruning and scaling data · 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
compress passages to 8 vectors instead of full textRankMistral full-text · TREC DL19nDCG@10: 0.7196 -> 0.7905Improved
compress passages to 8 vectors instead of full textRankMistral full-text · TREC DL19processed tokens: 9635.1 -> 800.0Improved
use C2R vs ListT5-3BBEIR per-dataset · CovidnDCG@10: 0.8470 -> 0.8217Mixed
add compressor pre-training (ablation)C2R · SciFactnDCG@10: 0.6529 -> 0.7276Improved
add SFT stage 1 (coarse alignment) (ablation)C2R · TREC DL20nDCG@10: 0.5516 -> 0.7791Improved
add SFT stage 2 (fine refinement) (ablation)C2R · TREC DL20nDCG@10: 0.5973 -> 0.7791Improved
remove SFT entirely (ablation)C2R · DL20, News, Robust04, SciFactnDCG@10Hurt
reduce compressed length from 32 to 8 (ablation)C2R · TREC DL19nDCG@10: 0.7382 -> 0.7905Improved
reduce compressed length from 32 to 8 (ablation)C2R · NewsnDCG@10: 0.4990 -> 0.5623Improved

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