Research map / Vision-language and multimodal / Multimodal LLMs
Instruction Tuning-free Visual Token Complement for Multimodal LLMs
Dongsheng Wang, Jiequan Cui, Miaoge Li, Wang Lin, Bo Chen, Hanwang Zhang
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add VTC complementary visual tokens to the frozen MLLM | BLIP2 · LVLM-eHub Visdial | score: 10.73 -> 77.20 | Improved |
| add VTC tokens to MiniGPT-4 | MiniGPT-4 · LVLM-eHub ScienceQA | accuracy: 25.43 -> 37.85 | Improved |
| add VTC tokens to InstructBLIP | InstructBLIP · LVLM-eHub ScienceQA | accuracy: 46.26 -> 46.21 | No clear effect |
| add VTC tokens to BLIP2 | BLIP2 · MME Code Reasoning | score: 75.00 -> 70.50 | Hurt |
| add VTC tokens after LoRA fine-tuning the LLM | InstructBLIP · MME Perception | score: 121.42 -> 124.31 | Improved |
| add 32 extra learnable visual queries (64 total) instead of VTC (ablation) | BLIP2 · MME Perception | score: 133.79 -> 133.95 | No clear effect |
| reuse VTC module for two iterations at test time instead of one (ablation) | VTC · OKVQA, IconQA, GQA | normalized VQA | Improved |
| increase iterations beyond two (ablation) | VTC · OKVQA, IconQA, GQA | normalized VQA | No 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.
- BLIP2 — BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation (ICML 2022) · better on 3 of 4 shared results
- MiniGPT-4 — MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models (ICLR 2024) · better on 1 of 1 shared result
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
- TokenSwap: Benchmarking and Reducing the Modality Gap in Multimodal LLMs (NeurIPS 2026 · similarity 0.92)
- Accelerating Pre-training of Multimodal LLMs via Chain-of-Sight (NeurIPS 2024 · similarity 0.91)
- Learning to Instruct for Visual Instruction Tuning (NeurIPS 2025 · similarity 0.90)
- VILA: On Pre-training for Visual Language Models (CVPR 2024 · similarity 0.90)
- Accelerating Multimodal Large Language Models via Dynamic Visual-Token Exit and the Empirical Findings (NeurIPS 2025 · similarity 0.90)
- Seeing is Understanding: Unlocking Causal Attention into Modality-Mutual Attention for Multimodal LLMs (ICML 2026 · similarity 0.90)
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.
