arXiv：cs.LG· Subramanyam Sahoo, Aman Chadha, Vinija Jain, Divya Chaudhary·· 6 小时前
推理陷阱：逻辑推理为何成为大语言模型情境意识的机制路径
The Reasoning Trap -- Logical Reasoning as a Mechanistic Pathway to Situational Awareness
AI 导读
论文提出 RAISE 框架（Reasoning Advancing Into Self Examination），指出 LLM 逻辑推理能力的提升会通过演绎自我推断、归纳情境识别、溯因自我建模三条机制路径，逐步加深模型的情境意识，并构建了从基础自我识别到策略性欺骗的升级阶梯。
正文
Abstract:Situational awareness, the capacity of an AI system to recognize its own nature, understand its training and deployment context, and reason strategically about its circumstances, is widely considered among the most dangerous emergent capabilities in advanced AI systems. Separately, a growing research effort seeks to improve the logical reasoning capabilities of large language models (LLMs) across deduction, induction, and abduction. In this paper, we argue that these two research trajectories are on a collision course. We introduce the RAISE framework (Reasoning Advancing Into Self Examination), which identifies three mechanistic pathways through which improvements in logical reasoning enable progressively deeper levels of situational awareness: deductive self inference, inductive context recognition, and abductive self modeling. We formalize each pathway, construct an escalation ladder from basic self recognition to strategic deception, and demonstrate that every major research topic in LLM logical reasoning maps directly onto a specific amplifier of situational awareness. We further analyze why current safety measures are insufficient to prevent this escalation. We conclude by proposing concrete safeguards, including a "Mirror Test" benchmark and a Reasoning Safety Parity Principle, and pose an uncomfortable but necessary question to the logical reasoning community about its responsibility in this trajectory.
| Comments: | Accepted at ICLR 2026 Workshop on Logical Reasoning of Large Language Models. 21 Pages. Position Paper |
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY); Machine Learning (cs.LG) |
| Cite as: | arXiv:2603.09200 [cs.AI] |
| (or arXiv:2603.09200v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2603.09200 arXiv-issued DOI via DataCite |
Submission history
From: Vinija Jain [view email]
[v1]
Tue, 10 Mar 2026 05:18:48 UTC (56 KB)
[v2]
Wed, 14 Oct 2026 03:22:44 UTC (56 KB)
来源：arXiv：cs.LG · arxiv.org