arXiv:2601.22139v2 Announce Type: replace 
Abstract: Reasoning-oriented Large Language Models (LLMs) have achieved remarkable progress with Chain-of-Thought (CoT) prompting, yet they remain fundamentally limited by a \emph{blind self-thinking} paradigm: performing extensive internal reasoning even when critical information is missing or ambiguous. We propose Proactive Interactive Reasoning (PIR), a new reasoning paradigm that transforms LLMs from passive solvers into proactive inquirers that interleave reasoning with clarification. Unlike existing search- or tool-based frameworks that primarily address knowledge uncertainty by querying external environments, PIR targets premise- and intent-level uncertainty through direct interaction with the user. PIR is implemented via two core components: (1) an uncertainty-aware supervised fine-tuning procedure that equips models with interactive reasoning capability, and (2) a user-simulator-based policy optimization framework driven by a composite reward that aligns model behavior with user intent. Extensive experiments on mathematical reasoning, code generation, and document editing demonstrate that PIR consistently outperforms strong baselines, achieving up to 32.70\% higher accuracy, 22.90\% higher pass rate, and 41.36 BLEU improvement, while reducing nearly half of the reasoning computation and unnecessary interaction turns. Further reliability evaluations on factual knowledge, question answering, and missing-premise scenarios confirm the strong generalization and robustness of PIR. Model and code are publicly available at: \href{https://github.com/SUAT-AIRI/Proactive-Interactive-R1}

أدت التطورات الأخيرة في نماذج اللغة الموجهة نحو التفكير (LLMs) إلى تقديم مفهوم التفكير التفاعلي الاستباقي (PIR)، وهو تحول في النموذج من الحلول السلبية إلى الاستفسارات الاستباقية، مما يمكّن النماذج من التفاعل مع المستخدمين للحصول على توضيحات أثناء عمليات التفكير. يتناول هذا النهج قيود التحفيز التقليدي لسلسلة التفكير من خلال دمج تفاعل المستخدم لحل حالات عدم اليقين.

Los recientes avances en modelos de lenguaje orientados al razonamiento (LLMs) han llevado a la introducción del Razonamiento Interactivo Proactivo (PIR), un cambio de paradigma que transforma la resolución pasiva de problemas en indagación proactiva, permitiendo que los modelos interactúen con los usuarios para aclaraciones durante los procesos de razonamiento. Este enfoque aborda las limitaciones del prompting de Cadena de Pensamiento tradicional al incorporar la interacción del usuario para resolver incertidumbres.

Des avancées récentes dans les modèles de langage orientés vers le raisonnement (LLMs) ont conduit à l'introduction du Raisonnement Interactif Proactif (PIR), un changement de paradigme passant de la résolution passive de problèmes à l'enquête proactive, permettant aux modèles d'engager les utilisateurs pour des clarifications lors des processus de raisonnement. Cette approche aborde les limitations de l'invite de chaîne de pensée traditionnelle en intégrant l'interaction utilisateur pour résoudre les incertitudes.

Recent advancements in reasoning-oriented Large Language Models (LLMs) have led to the introduction of Proactive Interactive Reasoning (PIR), a paradigm shift from passive problem-solving to proactive inquiry, enabling models to engage users for clarification during reasoning processes. This approach addresses the limitations of traditional Chain-of-Thought prompting by incorporating user interaction to resolve uncertainties.