AI Paper Atlas

Research map / Vision-language and multimodal / Visual question answering

When Robots Should Say ''I Don't Know'': Benchmarking Abstention in Embodied Question Answering

Tao Wu, Chuhao Zhou, Guangyu Zhao, Haozhi Cao, Yewen Pu, Jianfei Yang

CVPR 2026 · Vision-language and multimodal › Visual question answering · approach: Open-domain multimodal QA · 7 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 coarse abstention prompt (ablation)Qwen2.5-VL-7B · EM-EQAabstention recall / precision: recall 28.61 -> 75.86; precision 76.85 -> 56.84Mixed
add fine abstention prompt (ablation)HM-EQA agent · A-EQAabstention recall (%): 27.17 -> 99.46 (precision 56.82 -> 49.86)Mixed
enable thinking mode (ablation)Qwen3-VL-32B · EM-EQAabstention recall (%): 31.30 -> 21.70Hurt
enable thinking mode (ablation)Qwen3-VL-8B · EM-EQAabstention recall (%): 33.92 -> 27.02Hurt
enable thinking mode (ablation)Qwen3-VL-2B · EM-EQAabstention recall (%): 23.59 -> 25.55No clear effect
enable thinking mode (ablation)Gemini-2.5-Pro · EM-EQAabstention recall (%): 44.13 -> 43.09No clear effect
supervised fine-tuning on abstention (ablation)Qwen2.5-VL-7B · EM-EQAabstention recall (%): 26.94 -> 83.27Improved
replace visual inputs with random ones in SFT (ablation)7B-SFT · EM-EQAabstention recall (%): 83.27 -> 83.27No clear effect
remove visual input entirely (text-only SFT) (ablation)7B-SFT · EM-EQAabstention recall (%): 83.27 -> 86.12No clear effect
scale model within Qwen families (ablation)Qwen2.5-VL / Qwen3-VL · AbstainEQAabstention recall: e.g. 4B 27.32 vs 7B 28.61 across familiesMixed
ambiguous (abstention) queries instead of answerable ones3D-Mem · A-EQAsuccess rate (%): 77.17 -> 61.41Hurt

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