Show and Tell: Prompt Strategies for Style Control in Multi-Turn LLM Code Generation
This work addresses the challenge of making LLM-generated code more human-like in style for developers, though it is incremental as it focuses on a specific two-turn workflow.
The study tackled the problem of controlling code style in multi-turn LLM code generation, finding that combined prompts achieved the strongest initial compression (reducing verbosity) and greatest expansion discipline (maintaining style during revisions), while instruction-based prompts showed large initial effects and moderate discipline, and example-based prompts had modest initial effects with no discipline.
Language models generate functionally correct code that tends toward excessive verbosity, with elaborate documentation and defensive patterns that diverge from human baselines. Two prompting mechanisms have emerged for stylistic control: instruction based prompts that articulate abstract directives, and example based prompts that provide concrete code demonstrations. The core problem is whether stylistic constraints persist when models enhance initial implementations with additional features while maintaining high functional accuracy. Here we show that instruction-based, example-based, and combined prompts produce distinct patterns of initial control and expansion discipline over one enhancement turn. We manipulated system prompts across four conditions in a paired two-turn protocol where models first generated solutions to an intermediate Python task, then revised their code under general improvement directives, holding the user task fixed (N = 160 paired programs). Combined prompts produced the strongest initial compression and greatest expansion discipline. Instructions showed large initial effects and moderate expansion discipline. Examples showed modest initial effects with no expansion discipline. These results show that initial prompt effectiveness and expansion discipline are separate aspects of prompt design, and that combined approaches provide the most stable stylistic control in this two-turn workflow.