paper-with-me

홈 › Papers

Diagnosing Knowledge Conflict in Multimodal Long-Chain Reasoning

2026-02-16 · Jing Tang, Kun Wang, Haolang Lu, Hongjin Chen, KaiTao Chen, Zhongxiang Sun, Qiankun Li, Lingjuan Lyu, Guoshun Nan, Zhigang Zeng arxiv

Multimodal large language models (MLLMs) in long chain-of-thought reasoning often fail when different knowledge sources provide conflicting signals. We formalize these failures under a unified notion of knowledge conflict, distinguishing input-level objective conflict from process-level effective conflict. Through probing internal representations, we reveal that: (I) Linear Separability: different conflict types are explicitly encoded as linearly separable features rather than entangled; (II) Depth Localization: conflict signals concentrate in mid-to-late layers, indicating a distinct processing stage for conflict encoding; (III) Hierarchical Consistency: aggregating noisy token-level signals along trajectories robustly recovers input-level conflict types; and (IV) Directional Asymmetry: reinforcing the model's implicit source preference under conflict is far easier than enforcing the opposite source. Our findings provide a mechanism-level view of multimodal reasoning under knowledge conflict and enable principled diagnosis and control of long-CoT failures.

📄 PDF Abstract BibTeX arXiv:2602.14518

Code (0)

등록된 구현이 없습니다.

Tasks

Multimodal Reasoning

Similar Papers 제목 키워드 기반

BRIDGE: Benchmark for multi-hop Reasoning In long multimodal Documents with Grounded Evidence

2026-03-09 · Biao Xiang, Soyeon Caren Han, Yihao Ding arxiv

Multi-hop question answering (QA) is widely used to evaluate the reasoning capabilities of large language models, yet most benchmarks focus on final answer correctness and overlook intermediate reasoning, especially in l…

Multi-hop Question Answering

CrossCheck-Bench: Diagnosing Compositional Failures in Multimodal Conflict Resolution

2025-11-19 · Baoliang Tian, Yuxuan Si, Jilong Wang, Lingyao Li 외 arxiv

Multimodal Large Language Models are primarily trained and evaluated on aligned image-text pairs, which leaves their ability to detect and resolve real-world inconsistencies largely unexplored. In open-domain application…

Multimodal Reasoning

Benchmarking Multimodal Knowledge Conflict for Large Multimodal Models

2025-05-26 · Yifan Jia, Kailin Jiang, Yuyang Liang, Qihan Ren 외

Large Multimodal Models(LMMs) face notable challenges when encountering multimodal knowledge conflicts, particularly under retrieval-augmented generation(RAG) frameworks where the contextual information from external sou…

BenchmarkingRAGRetrieval-augmented Generation

SCoRE: Benchmarking Long-Chain Reasoning in Commonsense Scenarios

2025-03-08 · Weidong Zhan, Yue Wang, Nan Hu, Liming Xiao 외

Currently, long-chain reasoning remains a key challenge for large language models (LLMs) because natural texts lack sufficient explicit reasoning data. However, existing benchmarks suffer from limitations such as narrow …

BenchmarkingDiagnosticLogical ReasoningMultiple-choice

Getting Sick After Seeing a Doctor? Diagnosing and Mitigating Knowledge Conflicts in Event Temporal Reasoning

2023-05-24 · Tianqing Fang, Zhaowei Wang, Wenxuan Zhou, Hongming Zhang 외

Event temporal reasoning aims at identifying the temporal relations between two or more events from narratives. However, knowledge conflicts arise when there is a mismatch between the actual temporal relations of events …

counterfactualData AugmentationHallucinationIn-Context Learning