Let's Think in Two Steps: Mitigating Agreement Bias in MLLMs with Self-Grounded Verification

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Andrade, Moises, Cha, Joonhyuk, Ho, Brandon, Srihari, Vriksha, Yadav, Karmesh, Kira, Zsolt
Format:	Preprint
Published:	2025
Subjects:	Artificial Intelligence Computation and Language Machine Learning Multiagent Systems Robotics
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2507.11662