Intention Chain-of-Thought Prompting with Dynamic Routing for Code Generation

2025-12-15

AI Summary

Existing chain-of-thought (CoT) prompting for code generation suffers from two key limitations: (1) reasoning redundancy—overly complex reasoning for simple tasks—and (2) inadequate intent modeling—neglecting algorithmic logic and computational complexity objectives. To address these, we propose RoutingGen, the first dynamic routing framework enabling *on-demand reasoning*: it adaptively selects between few-shot prompting and our novel Intent Chain-of-Thought (ICoT) based on task difficulty. ICoT explicitly models high-level intents—including core algorithms, time complexity constraints, and problem semantics—and incorporates a cognition-aware, dynamically difficulty-perceived routing mechanism grounded in cognitive economy principles. Evaluated across three large language models and six diverse benchmarks, RoutingGen achieves state-of-the-art performance while reducing average token consumption by 46.37%. Notably, ICoT significantly outperforms six mainstream prompting methods on high-difficulty tasks.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Natural Language Processing: Generation
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Large language models (LLMs) exhibit strong generative capabilities and have shown great potential in code generation. Existing chain-of-thought (CoT) prompting methods enhance model reasoning by eliciting intermediate steps, but suffer from two major limitations: First, their uniform application tends to induce overthinking on simple tasks. Second, they lack intention abstraction in code generation, such as explicitly modeling core algorithmic design and efficiency, leading models to focus on surface-level structures while neglecting the global problem objective. Inspired by the cognitive economy principle of engaging structured reasoning only when necessary to conserve cognitive resources, we propose RoutingGen, a novel difficulty-aware routing framework that dynamically adapts prompting strategies for code generation. For simple tasks, it adopts few-shot prompting; for more complex ones, it invokes a structured reasoning strategy, termed Intention Chain-of-Thought (ICoT), which we introduce to guide the model in capturing task intention, such as the core algorithmic logic and its time complexity. Experiments across three models and six standard code generation benchmarks show that RoutingGen achieves state-of-the-art performance in most settings, while reducing total token usage by 46.37% on average across settings. Furthermore, ICoT outperforms six existing prompting baselines on challenging benchmarks.

Problem

• Dynamically adapts prompting strategies for code generation
• Reduces overthinking on simple tasks via difficulty-aware routing
• Captures task intention like algorithmic logic and efficiency

Innovation

• Dynamic routing framework adapts prompting strategies based on task difficulty
• Intention Chain-of-Thought guides model to capture core algorithmic logic
• Reduces token usage while achieving state-of-the-art performance