AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Are They the Same? Exploring Visual Correspondence Shortcomings of Multimodal LLMs

Yikang Zhou, Tao Zhang, Shilin Xu, Shihao Chen, Qianyu Zhou, Yunhai Tong, Shunping Ji, Jiangning Zhang, Lu Qi, Xiangtai Li

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Instruction-tuned multimodal assistants · 15 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 MMVM SFT data with LLaVA SFT data (ablation)InternVL2-4B · MMVM benchmark (ablation, 30% data)Overall accuracy: 17.62 -> 32.38Improved
add object-level contrastive learning to the MLLM visual encoder only (ablation)InternVL2-4B + Data · MMVM benchmark (ablation, 30% data)Overall accuracy: 32.38 -> 34.05Improved
add RADIO vision expert without contrastive alignment (ablation)InternVL2-4B + Data · MMVM benchmark (ablation, 30% data)Overall accuracy: 32.38 -> 32.25No clear effect
add vision expert with OCL pre-training of its adapter (ablation)InternVL2-4B + Data + OCL · MMVM benchmark (ablation, 30% data)Overall accuracy: 34.05 -> 40.45Improved
add instruction augmentation with object-level representations (ablation)InternVL2-4B + Data + OCL + VE · MMVM benchmark (ablation, 30% data)Overall accuracy: 40.45 -> 45.83Improved
pre-train adapter with region-region contrastive loss instead of image-text (ablation)CoLVA · MMVM benchmark (adapter pre-training ablation)Overall accuracy: 33.64 -> 40.45Improved
pre-train adapter with region-region contrastive loss instead of region-text (ablation)CoLVA · MMVM benchmark (adapter pre-training ablation)Overall accuracy: 30.93 -> 40.45Improved
use RADIO instead of DINOv2 as vision expert (ablation)CoLVA · MMVM benchmark (vision expert alternatives)Overall accuracy: 40.34 -> 45.83Improved
use RADIO instead of ConvNext-L as vision expert (ablation)CoLVA · MMVM benchmark (vision expert alternatives)Overall accuracy: 39.80 -> 45.83Improved
integrate CoLVA into LLaVA1.5-7B without multi-image trainingLLaVA1.5-7B · MMVM benchmark (generalization to other MLLMs)Overall accuracy: 14.64 -> 36.56Improved
integrate CoLVA into Qwen2VL-2BQwen2VL-2B · MMVM benchmark (generalization to other MLLMs)Overall accuracy: 15.69 -> 47.48Improved
integrate CoLVA into InternVL2-4B on MMMU ValInternVL2-4B · MMMU ValOverall accuracy: 47.56 -> 44.11Hurt

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