AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

VIVAS: Vitalizing Visual Perception in VLM Pre-training via Vision-language Unified Autoregressive Supervision

Zhehan Kan, Yubo Zhu, Xinghua Jiang, Zhixiang Wei, Shifeng Liu, Wei Tong, Sheng Zhong, Qingmin Liao, Wenming Yang, Xin Li, Yinsong Liu, Deqiang Jiang and 1 more

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Instruction-tuned multimodal assistants · 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 dense-structural-semantic visual supervision in pre-training (ablation)VLM (0.5T) · VisuLogic (0.5T, 5-shot)accuracy: 22.2 -> 24.4Improved
add dense-structural-semantic visual supervision in pre-training (ablation)VLM (0.5T) · RealWorldQA (0.5T, 5-shot)accuracy: 43.0 -> 46.5Improved
add text-decoupled (VQ-VAE) visual vocabulary supervision (ablation)VLM (0.5T) · MMBench-EN (0.5T, 5-shot)accuracy: 53.0 -> 51.2Hurt
add vision-language semantic (SigLIP2) vocabulary supervision (ablation)VLM (0.5T) · MMBench-EN (0.5T, 5-shot)accuracy: 53.0 -> 53.6No clear effect
add visual supervision in post-training only (ablation)VIVAS 4B · VisuLogicaccuracy: 23.1 -> 23.3No clear effect
add visual supervision in pre-training only (ablation)VIVAS 4B · VisuLogicaccuracy: 23.1 -> 25.7Improved
also add visual supervision in post-training on top of pre-training (ablation)VIVAS 4B · VisuLogicaccuracy: 25.7 -> 23.6Hurt
add L1 reconstruction loss to codebook training (ablation)VIVAS tokenizer · VisuLogic (0.5T, 5-shot)accuracy: 24.4 -> 23.7Hurt
increase codebook size from 50k to 150k (ablation)VIVAS tokenizer · VisuLogic (0.5T, 5-shot)accuracy: 23.6 -> 24.4Improved
increase codebook size from 150k to 200k (ablation)VIVAS tokenizer · VisuLogic (0.5T, 5-shot)accuracy: 24.4 -> 24.2Hurt
increase codebook dim from 64 to 768 (ablation)VIVAS tokenizer · VisuLogic (0.5T, 5-shot)accuracy: 22.7 -> 24.4Improved
replace visual supervision with combined SigLIP2+DINOv3 input encoder (ablation)VLM (0.5T) · RefCOCO val (0.5T, 5-shot)accuracy: 62.1 -> 60.9Hurt

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