AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Calibrating MLLM-as-a-judge via Multimodal Bayesian Prompt Ensembles

Eric Slyman, Mehrab Tanjim, Kushal Kafle, Stefan Lee

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Unified any-modality language models · 3 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 image-cluster-conditioned prompt weights (MMB vs BPE)Bayesian prompt ensemble · HPSv2 pairwise (10 prompts, 20 validation samples)NLL: .547 -> .430Improved
add image-cluster-conditioned prompt weights (MMB vs BPE)Bayesian prompt ensemble · HPSv2 pairwise (10 prompts, 20 validation samples)F1: .753 -> .778Improved
add image-cluster-conditioned prompt weights (MMB vs BPE)Bayesian prompt ensemble · HPSv2 pairwise (10 prompts, 20 validation samples)Brier score: .147 -> .135Improved
add image-cluster-conditioned prompt weights (MMB vs BPE)Bayesian prompt ensemble · HPSv2 pairwise (10 prompts, 20 validation samples)ECE: .114 -> .091Improved
use MMB instead of unweighted averagingprompt ensemble · HPSv2 pairwise (10 prompts, 20 validation samples)F1: .764 -> .778Improved
use MMB instead of unweighted averagingprompt ensemble · HPSv2 pairwise (10 prompts, 20 validation samples)ECE: .142 -> .091Improved
use ensemble (BPE vs Avg.) weightingprompt ensemble · HPSv2 pairwise (10 prompts, 20 validation samples)F1: .764 -> .753Hurt
use MMB instead of the best single promptsingle prompt · HPSv2 pairwise (10 prompts, 20 validation samples)accuracy: .812 -> .820Improved
increase cluster count K from 4 to 16 (ablation)MMB · HPSv2 pairwise (10 prompts, 20 validation samples)NLL: .432 -> .430Improved
increase cluster count K from 4 to 64 (ablation)MMB · HPSv2 pairwise (10 prompts, 20 validation samples)ECE: .090 -> .091No clear effect
use MMB instead of unweighted averagingprompt ensemble · MJBench-Bias synthetic no-preferenceaverage confidence: .656 -> .705Hurt
use MMB instead of BPEprompt ensemble · MJBench-Bias synthetic no-preferenceaverage confidence: .715 -> .705Improved

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