AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

MM-IFEngine: Towards Multimodal Instruction Following

Shengyuan Ding, Shenxi Wu, Xiangyu Zhao, Yuhang Zang, Haodong Duan, Xiaoyi Dong, Pan Zhang, Yuhang Cao, Dahua Lin, Jiaqi Wang

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Visual instruction tuning · 36 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
fine-tune with MM-IFDPO-23k (DPO)Qwen2-VL-7B-Instruct · MM-IFEvalaccuracy: 42.0 -> 52.2Improved
fine-tune with MM-IFInstruct-23k (SFT)Qwen2-VL-7B-Instruct · MM-IFEval + MIA + IFEval averageaverage: 56.6 -> 64.2Improved
fine-tune with MM-IFDPO-23k (DPO)LLaVA-NeXT-Llama3-8B · MM-IFEval + MIA + IFEval averageaverage: 57.9 -> 69.5Improved
fine-tune with MM-IFDPO-23k (DPO)LLaVA-NeXT-Llama3-8B · IFEval (text-only)accuracy (avg): 50.7 -> 69.1Improved
use DPO instead of SFT on the MM-IF dataLLaVA-NeXT-Llama3-8B · IFEval (text-only)accuracy (avg): 56.3 -> 69.1Improved
fine-tune with MM-IFDPO-23k (DPO)Qwen2-VL-7B-Instruct · VQA benchmarks average (8 benchmarks)average: 72.3 -> 72.4No clear effect
fine-tune with MM-IFInstruct-23k (SFT)Qwen2-VL-7B-Instruct · VQA benchmarks average (8 benchmarks)average: 72.3 -> 70.7Hurt
fine-tune with MM-IFDPO-23k (DPO)LLaVA-NeXT-Llama3-8B · MM-IFEval perception-levelaccuracy: 21.0 -> 21.0No clear effect
remove 100% of constraints for rejected DPO responses instead of 33% (ablation)Qwen2-VL-7B-Instruct · MM-IFEval + MIA + IFEval averageaverage: 65.8 -> 66.7Improved
remove 100% of constraints for rejected DPO responses instead of 33% (ablation)LLaVA-NeXT-Llama3-8B · MM-IFEval + MIA + IFEval averageaverage: 67.3 -> 69.5Improved
generate DPO rejected responses with the image present instead of without image (ablation)Qwen2-VL-7B-Instruct · MM-IFEval + MIA + IFEval averageaverage: 63.4 -> 66.7Improved
generate DPO rejected responses with the image present instead of without image (ablation)LLaVA-NeXT-Llama3-8B · MM-IFEvalaccuracy: 44.7 -> 49.3Improved

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