AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

ProactiveBench: Benchmarking Proactiveness in Multimodal Large Language Models

Thomas De Min, Subhankar Roy, Stéphane Lathuilière, Elisa Ricci, Massimiliano Mancini

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Unified any-modality language models · 6 citations (OpenAlex)

Read the paper

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
scale model size from InternVL3-1B to 8B (ablation)InternVL3 · ProactiveBench MCQAaccuracy: 27.1% vs 12.7%Hurt
use newer larger LLaVA-OV-72B instead of LLaVA-1.5-7B (ablation)LLaVA · ProactiveBench MCQAaccuracy: 24.8% vs 13.0%Hurt
use Vicuna instead of Mistral LLM (ablation)LLaVA-NeXT-7B · ProactiveBench MCQAaccuracy: 4.5% -> 19.3%Improved
add proactive hints to the prompt (ablation)MLLMs · ProactiveBench MCQAproactive suggestion rate: +1.9Improved
add proactive hints to the prompt (ablation)MLLMs · ProactiveBench MCQAaccuracy: 25.8% (+8.3%)Mixed
condition on conversation history (ablation)MLLMs · ProactiveBench MCQAaccuracy: -7% accuracy; psr 0.5 -> 1.8Hurt
add 1 or 3 few-shot in-context examples (ablation)MLLMs · ProactiveBench RODaccuracy: 6.7% with one sample, 12.0% with threeHurt
replace valid proactive options with random invalid ones (ablation)LLaVA-NeXT-Vicuna-7B · ProactiveBench MCQAproactive action probability: 37% -> 49%No clear effect
GRPO post-training with proactive reward r_p=0.5LLaVA-NeXT-Mistral-7B · ProactiveBench MCQAaccuracy: 4.5 -> 40.4Improved
GRPO post-training with proactive reward r_p=0.75Qwen2.5-VL-3B · ProactiveBench MCQAaccuracy: 11.0 -> 38.6Improved
raise proactive reward to r_p=1.0 (ablation)Qwen2.5-VL-3B · ProactiveBench MCQAaccuracy: 38.6 -> 17.2Hurt
raise proactive reward to r_p=1.0 equal to correct reward (ablation)Qwen2.5-VL-3B · ProactiveBench MCQAproactive suggestion rate: 0.4 to 15.1Hurt

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