AI Paper Atlas

Research map / Video / Video-language understanding

V-LynX: Token Interface Alignment for Video+X LLMs

Jungin Park, Jiyoung Lee, Kwanghoon Sohn

ICML 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 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
add attention response alignment (ablation)V-LynX-0.5B · ScanQACIDEr: 81.0 -> 87.1Improved
add distribution regularization (ablation)V-LynX-0.5B · ScanQACIDEr: 86.2 -> 87.1Improved
add interface alignment stage before instruction tuning (ablation)V-LynX-0.5B · ScanQACIDEr: 77.3 -> 87.1Improved
add attention response alignment (ablation)V-LynX-0.5B · ScanQAEM@1: 23.5 -> 26.4Improved
raise vision-encoder LoRA rank from 8 to 64 (ablation)V-LynX-0.5B · ScanQACIDEr: 86.1 -> 87.1Improved
use audio-related videos instead of all as reference set (ablation)V-LynX-0.5B · ScanQACIDEr: 87.7 vs 87.1No clear effect
use only 563 3D clips as reference set (ablation)V-LynX-0.5B · ScanQACIDEr: 87.8 vs 87.1No clear effect
replace PAVE with V-LynXLLaVA-OV-7B · AVSDCIDEr: +10.1 CIDErImproved
replace PAVE with V-LynXLLaVA-OV-7B · ScanQA/SQA3Dparameters: 195.0M vs. 475.0MImproved
apply V-LynX to Qwen2.5-VL-3BQwen2.5-VL-3B · SQA3DEM@1: 15.1 -> 59.7Improved
apply V-LynX to InternVL-2.5-4BInternVL-2.5-4B · SQA3DEM@1: 44.0 -> 61.1Improved
replace PAVE-0.5B with V-LynX-0.5BLLaVA-OV-0.5B · VideoMMEaccuracy: +6.8%Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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