The Imitation Game: When LLMs Learn to Reason Like Programs via Code-Centric Reasoning Data Synthesis

2026-09-13

AI Summary

为解决大语言模型在自然语言精细推理上的不足，提出MIMIC框架，通过代码合成数据进行训练，提升模型的广泛推理能力。

Abstract

Large Language Models (LLMs) excel at programming tasks but frequently fail at deterministic, fine-grained reasoning in natural language, relying heavily on semantic approximations rather than robust symbolic execution. To bridge this gap, we propose MIMIC, a framework that leverages executable code as a rigorous medium for reasoning data synthesis. MIMIC fundamentally transforms algorithms into verifiable reasoning trajectories through narrative fusion, code-guided test synthesis, and dynamic code instrumentation. Crucially, these explicit intermediate execution states naturally form a Code-Instrumented Reward (CIR), providing dense, high-fidelity process supervision for reinforcement learning without external reward models. Extensive evaluations reveal that models trained via SFT and GRPO on our synthesized dataset achieve substantial, consistent gains. Our method significantly elevates accuracy across general reasoning, complex mathematical benchmarks, and fine-grained deterministic tasks, demonstrating that the procedural rigor of executable code can effectively unlock and enhance the generalized reasoning capabilities of LLMs. Our code and data are available at https://github.com/zjy1298/MIMIC.

Problem

• Large Language Models
• deterministic reasoning
• fine-grained reasoning
• symbolic execution
• natural language

Innovation

• MIMIC
• Code-Instrumented Reward (CIR)
• code-centric reasoning
• reinforcement learning
• executable code