AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

The Cost of Reasoning: Chain-of-Thought Induces Overconfidence in Vision-Language Models

Robert Welch, Emir Konuk, Kevin Smith

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Chain-of-thought multimodal reasoning · 4 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 CoT promptingMSP uncertainty, Qwen3-VL-8B-IT · Oxford-IIIT PetPRR: 0.710 -> 0.231Hurt
add CoT promptingMSP uncertainty, Qwen3-VL-8B-IT · MMMU-Pro-VisionPRR: 0.450 -> 0.123Hurt
add CoT promptingConsistency, Qwen3-VL-8B-IT · MathVistaPRR: 0.345 -> 0.684Improved
add CoT promptingConsistency, Qwen3-VL-8B-IT · MMMU-Pro-VisionPRR: 0.107 -> 0.434Improved
add CoT promptingQwen3-VL-8B-IT · MathVistaaccuracy: 63.6 -> 70.4Improved
add CoT promptingQwen3-VL-8B-IT · OK-VQAaccuracy: 55.9 -> 52.1Hurt
use reasoning-trained Thinking model instead of InstructMTE, Qwen3-VL-8B · Oxford-IIIT PetPRR: 0.735 -> 0.086Hurt
use reasoning-trained Thinking model instead of InstructMTE, Qwen3-VL-8B · MathVistaPRR: 0.413 -> -0.056Hurt
mask answer mentions in CoT trace (ablation)MSP, Qwen3-VL-8B-IT CoT · OK-VQAPRR: 0.141 -> 0.475Improved
mask answer mentions in CoT trace (ablation)MTE, Qwen3-VL-8B-IT CoT · OK-VQAPRR: 0.071 -> 0.329Improved
randomly mask equal number of reasoning tokens (ablation)MSP, Qwen3-VL-8B-IT CoT · OK-VQAPRR: 0.141 -> 0.210Mixed
add reasoning length as explanation of confidence inflation (ablation)ATL estimates · all datasetspartial Spearman with lengthNo clear effect

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