Image de couverture de livre

When Modalities Conflict: How Unimodal Reasoning Uncertainty Governs Preference Dynamics in MLLMs

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Zhang, Zhuoran, Wang, Tengyue, Gong, Xilin, Shi, Yang, Wang, Haotian, Wang, Di, Hu, Lijie
Format:	Preprint
Publié:	2025
Sujets:	Artificial Intelligence
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Understanding the Dynamics of Demonstration Conflict in In-Context Learning
par: Jiao, Difan, et autres
Publié: (2026)

When Safe Unimodal Inputs Collide: Optimizing Reasoning Chains for Cross-Modal Safety in Multimodal Large Language Models
par: Cai, Wei, et autres
Publié: (2025)

Mechanistic Unveiling of Transformer Circuits: Self-Influence as a Key to Model Reasoning
par: Zhang, Lin, et autres
Publié: (2025)

MONICA: Real-Time Monitoring and Calibration of Chain-of-Thought Sycophancy in Large Reasoning Models
par: Hu, Jingyu, et autres
Publié: (2025)

STARS: Skill-Triggered Audit for Request-Conditioned Invocation Safety in Agent Systems
par: Zhang, Guijia, et autres
Publié: (2026)

Beyond Scalars: Evaluating and Understanding LLM Reasoning via Geometric Progress and Stability
par: Jiang, Xinyan, et autres
Publié: (2026)

MLLMs are Deeply Affected by Modality Bias
par: Zheng, Xu, et autres
Publié: (2025)

Locate-then-edit for Multi-hop Factual Recall under Knowledge Editing
par: Zhang, Zhuoran, et autres
Publié: (2024)

Adaptive Unimodal Regulation for Balanced Multimodal Information Acquisition
par: Huang, Chengxiang, et autres
Publié: (2025)

CODEMENV: Benchmarking Large Language Models on Code Migration
par: Cheng, Keyuan, et autres
Publié: (2025)

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification
par: Zhang, Lin, et autres
Publié: (2025)

Controlling Repetition in Protein Language Models
par: Zhang, Jiahao, et autres
Publié: (2026)

WebUncertainty: Dual-Level Uncertainty Driven Planning and Reasoning For Autonomous Web Agent
par: Zhang, Lingfeng, et autres
Publié: (2026)

Not Search, But Scan: Benchmarking MLLMs on Scan-Oriented Academic Paper Reasoning
par: Li, Rongjin, et autres
Publié: (2026)

D-CAT: Decoupled Cross-Attention Transfer between Sensor Modalities for Unimodal Inference
par: Daher, Leen, et autres
Publié: (2025)

OmniDPO: A Preference Optimization Framework to Address Omni-Modal Hallucination
par: Chen, Junzhe, et autres
Publié: (2025)

Can MLLMs Absorb Math Reasoning Abilities from LLMs as Free Lunch?
par: Hu, Yijie, et autres
Publié: (2025)

Rethinking the Text-Vision Reasoning Imbalance in MLLMs through the Lens of Training Recipes
par: Yao, Guanyu, et autres
Publié: (2025)

Tracking the Limits of Knowledge Propagation: How LLMs Fail at Multi-Step Reasoning with Conflicting Knowledge
par: Feng, Yiyang, et autres
Publié: (2026)

A Multi-Modal CNN-LSTM Framework with Multi-Head Attention and Focal Loss for Real-Time Elderly Fall Detection
par: Zhou, Lijie, et autres
Publié: (2026)

In-Run Data Shapley for Adam Optimizer
par: Ding, Meng, et autres
Publié: (2026)

CodePMP: Scalable Preference Model Pretraining for Large Language Model Reasoning
par: Yu, Huimu, et autres
Publié: (2024)

Visual Self-Fulfilling Alignment: Shaping Safety-Oriented Personas via Threat-Related Images
par: Yang, Qishun, et autres
Publié: (2026)

DynamicPO: Dynamic Preference Optimization for Recommendation
par: Hu, Xingyu, et autres
Publié: (2026)

Semi-supervised Concept Bottleneck Models
par: Hu, Lijie, et autres
Publié: (2024)

Do MLLMs Really Understand Space? A Mathematical Reasoning Evaluation
par: Lu, Shuo, et autres
Publié: (2026)

Predicting LLM Output Length via Entropy-Guided Representations
par: Xie, Huanyi, et autres
Publié: (2026)

LEGO-Puzzles: How Good Are MLLMs at Multi-Step Spatial Reasoning?
par: Tang, Kexian, et autres
Publié: (2025)

Blinded by Generated Contexts: How Language Models Merge Generated and Retrieved Contexts When Knowledge Conflicts?
par: Tan, Hexiang, et autres
Publié: (2024)

MMPareto: Boosting Multimodal Learning with Innocent Unimodal Assistance
par: Wei, Yake, et autres
Publié: (2024)

FairReason: Balancing Reasoning and Social Bias in MLLMs
par: Pan, Zhenyu, et autres
Publié: (2025)

Adaptive Multi-Subspace Representation Steering for Attribute Alignment in Large Language Models
par: Jiang, Xinyan, et autres
Publié: (2025)

Where and What: Reasoning Dynamic and Implicit Preferences in Situated Conversational Recommendation
par: Lin, Dongding, et autres
Publié: (2026)

Omni-Safety under Cross-Modality Conflict: Vulnerabilities, Dynamics Mechanisms and Efficient Alignment
par: Wang, Kun, et autres
Publié: (2026)

PAHQ: Accelerating Automated Circuit Discovery through Mixed-Precision Inference Optimization
par: Wang, Xinhai, et autres
Publié: (2025)

Scaling Spatial Reasoning in MLLMs through Programmatic Data Synthesis
par: Helu, Zhi, et autres
Publié: (2025)

Global Evolutionary Steering: Refining Activation Steering Control via Cross-Layer Consistency
par: Jiang, Xinyan, et autres
Publié: (2026)

StemBind: When MLLMs Get Lost Between Rules and Instances in Abstract Visual Reasoning
par: He, Xixiang, et autres
Publié: (2026)

SightSound-R1: Cross-Modal Reasoning Distillation from Vision to Audio Language Models
par: Wang, Qiaolin, et autres
Publié: (2025)

CSR-Bench: A Benchmark for Evaluating the Cross-modal Safety and Reliability of MLLMs
par: Liu, Yuxuan, et autres
Publié: (2026)