SeeingEye: Agentic Information Flow Unlocks Multimodal Reasoning in Text-only LLMs
Published in arXiv preprint arXiv:2510.25092; under review at EMNLP 2026, 2025
This work introduces an agentic information-flow approach that helps text-only LLMs perform multimodal reasoning.
Recommended citation: Weijia Zhang*, Zijia Liu*, Haoru Li*, Haoqi Chen*, and Jiaxuan You. (2025). "SeeingEye: Agentic Information Flow Unlocks Multimodal Reasoning in Text-only LLMs." arXiv preprint arXiv:2510.25092. Under review at EMNLP 2026.
Download Paper