AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Combating Textual Noise and Redundancy: Entropy-Aware Dense Visual Token Pruning

Xuehui Wang, Xuankun Yang, Wei Shen

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Attention-based visual token pruning · 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 global guidance to dense guidance (mu 0.0 -> 0.5) (ablation)EADP · LLaVA-1.5-7B, 128 tokens, 5-benchmark avgaverage accuracy: 67.7 -> 68.5Improved
add spatial smoothing (ablation)EADP · LLaVA-1.5-7B, 128 tokens, 5-benchmark avgaverage accuracy: 67.9 -> 68.5Improved
add score polarization (beta 0.0 -> 2.0) (ablation)EADP · LLaVA-1.5-7B, 128 tokens, 5-benchmark avgaverage accuracy: 67.8 -> 68.2Improved
discard high-entropy text tokens (q 0.9 -> 0.3) (ablation)EADP · LLaVA-1.5-7B, 128 tokens, 5-benchmark avgaverage accuracy: 67.6 -> 68.6Improved
use proposed aggregation instead of simple mean (ablation)EADP · LLaVA-1.5-7B, 128 tokens, 5-benchmark avgaverage accuracy: 68.0 -> 68.6Improved
replace entropy with central mass ratio or variance to measure dispersion (ablation)EADP · LLaVA-1.5-7B, 128 tokens, 5-benchmark avgaverage accuracy: 68.5 -> 68.4 (B), 68.5 (C)No clear effect
increase smoothing kernel from 3 to 5 or 7 (ablation)EADP · LLaVA-1.5-7B, 128 tokens, 5-benchmark avgaverage accuracy: 68.5 -> 68.5 (size=5), 68.4 (size=7)No clear effect
EADP vs DivPrune at 32 tokensEADP · LLaVA-1.5-7B, 32 tokensaverage accuracy: 58.4 -> 60.9Improved
EADP vs CDPruner on Qwen3-VL-8B at 256 tokensEADP · Qwen3-VL-8B DocVQA, 256 tokensaccuracy: 53.0 -> 62.8Improved
EADP vs DivPrune on Qwen2.5-VL at 512 tokensEADP · Qwen2.5-VL-7B, 512 tokensaverage score: 78.1 -> 78.0No clear effect
EADP vs DivPrune at 64x16 video tokensEADP · LLaVA-Video-7B, 64x16 tokensaverage accuracy: 54.0 -> 54.8Improved
EADP overhead vs HiPrune at 128 tokensEADP · LLaVA-1.5-7B, 128 tokensprefill time (ms): 99.4 -> 118.8Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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