SituatedThinker: Grounding LLM Reasoning with Real-World through Situated Thinking

ICLR 2026 Conference Submission13039 Authors

Keywords: large language models, agentic reinforcement learning, large language model reasoning, interfaces

TL;DR: We propose SituatedThinker to enable LLM reasoning within real-world environment.

Abstract: Recent advances in large language models (LLMs) demonstrate their impressive reasoning capabilities. However, the reasoning confined to internal parametric space limits LLMs' access to real-time information and understanding of the physical world. To overcome this constraint, we introduce SituatedThinker, a novel framework that enables LLMs to ground their reasoning in real-world contexts through situated thinking, which adaptively combines both internal knowledge and external information with predefined interfaces. By utilizing reinforcement learning, SituatedThinker incentivizes deliberate reasoning with the real world to acquire information and feedback, allowing LLMs to surpass their knowledge boundaries and enhance reasoning. Experimental results demonstrate significant performance improvements on multi-hop question-answering and mathematical reasoning benchmarks. Furthermore, SituatedThinker demonstrates strong performance on unseen tasks, such as KBQA, TableQA, and text-based games, showcasing the generalizable real-world grounded reasoning capability.

Primary Area: foundation or frontier models, including LLMs

Submission Number: 13039