AI Paper Atlas

Research map / Video / Video-language understanding

Unified Spatiotemporal Token Compression for Video-LLMs at Ultra-Low Retention

Junhao Du, Jialong Xue, Anqi Li, Jincheng Dai, Guo Lu

CVPR 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 4 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 similarity-based pruning to attention selection (ablation)attention-only selection · LLaVA-OV-7B 2% retention, 5 benchmarksAvg. score: 41.8 -> 49.7Improved
add clustering-based merging (ablation)attention + similarity · LLaVA-OV-7B 2% retention, 5 benchmarksAvg. score: 49.7 -> 50.4Improved
add attention scores (ablation)similarity + cluster · LLaVA-OV-7B 2% retention, 5 benchmarksAvg. score: 49.5 -> 50.4Improved
add similarity pruning (ablation)attention + cluster · LLaVA-OV-7B 2% retention, 5 benchmarksAvg. score: 45.8 -> 50.4Improved
add text-aware inner-LLM merging (ablation)unified spatiotemporal compression · LLaVA-OV-7B 2% retentionAvg. % of baseline: 89.5 -> 90.1Improved
add text-aware inner-LLM merging (ablation)unified spatiotemporal compression · LLaVA-OV-7B 5% retentionAvg. % of baseline: 94.9 -> 95.4Improved
add text-aware inner-LLM merging (ablation)unified spatiotemporal compression · LLaVA-OV-7B 1% retentionAvg. % of baseline: 84.2 -> 84.1No clear effect
use text-aware scoring instead of last-token attention (ablation)inner-LLM pruning · LLaVA-OV-7B 5% retentionAvg. % of baseline: 93.8 -> 95.4Improved
raise similarity threshold tau from 0.7 to 0.9 (ablation)Ours w/o M · LLaVA-OV-7B 2% retention, 5 benchmarksAvg. %: 89.5 -> 86.0Hurt
prune inside LLM at layer 18 instead of layer 2 (ablation)Ours · LLaVA-OV-7B 5 benchmarksAvg. %: 83.6 -> 90.1Improved
set cluster ratio 0.3 instead of 0.7 (ablation)Ours w/o M · LLaVA-OV-7B 2% retention, 5 benchmarksAvg. %: 87.6 -> 89.5Improved
set lambda 0.5 instead of 0 in text-aware merging (ablation)Ours · LLaVA-OV-7B 5 benchmarksAvg. %: 89.7 -> 90.1No clear effect

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