How Do LLMs and VLMs Understand Viewpoint Rotation Without Vision? An Interpretability Study
Zhen Yang, Ping Jian, Zhongbin Guo, Zuming Zhang, Chengzhi Li, Yonghong Deng, Xinyue Zhang, Wenpeng Lu
2026年份
摘要
Table!! Unknown Figure 1 : For textual viewpoint rotation understanding, human can easily imagine the spatial scenario and achieve 100% accuracy, while LLMs and VLMs struggle to output the correct answer.
问问这篇 Paper
智能体会读完全文。
Lune 把这篇 Paper 索引到了每一个公式,引用它的顶会 Paper 也一样。你提问,回答直接引用原文。
它引用的顶会 Paper24
- Chain-of-Thought Prompting Elicits Reasoning in Large Language ModelsJason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma 等NeurIPS 2022 · 被引用 22,562 次
- Measuring Massive Multitask Language UnderstandingDan Hendrycks, Collin Burns, Steven Basart, Andy Zou 等ICLR 2021 · 被引用 7,905 次
- Inference-Time Intervention: Eliciting Truthful Answers from a Language ModelKenneth Li, Oam Patel, Fernanda B. Viégas, Hanspeter Pfister 等NeurIPS 2023 · 被引用 1,549 次
- Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free LunchLe Yu, Bowen Yu, Haiyang Yu, Fei Huang 等ICML 2024 · 被引用 605 次
- DoLa: Decoding by Contrasting Layers Improves Factuality in Large Language ModelsYung-Sung Chuang, Yujia Xie, Hongyin Luo, Yoon Kim 等ICLR 2024 · 被引用 354 次
相关 Paper
- Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall SpacesJihan Yang, Shusheng Yang, Anjali W. Gupta, Rilyn Han 等CVPR 2025
- TopViewRS: Vision-Language Models as Top-View Spatial ReasonersChengzu Li, Caiqi Zhang, Han Zhou, Nigel Collier 等EMNLP 2024 · 被引用 5 次
- SpinBench: Perspective and Rotation as a Lens on Spatial Reasoning in VLMsYuyou Zhang, Radu Corcodel, Chiori Hori, Anoop Cherian 等ICLR 2026 · 被引用 11 次
- Spatial CAPTCHA: Generatively Benchmarking Spatial Reasoning for Human-Machine DifferentiationArina Kharlamova, Bowei He, Chen Ma, Xue LiuICLR 2026 · 被引用 3 次
- Benchmarking Large Vision-Language Models on CFMME: A Comprehensive Chinese Financial Multimodal Evaluation DatasetQian Chen, Xianyin Zhang, Yanzhi Liu, Lifan Guo 等ACL 2026
