Prompt Repetition Improves Non-Reasoning LLMs

This addresses a practical issue for users of LLMs in non-reasoning applications, though it is incremental as it builds on existing prompting methods.

The paper tackles the problem of improving performance for non-reasoning tasks in large language models by repeating the input prompt, finding that this simple technique enhances results for models like Gemini, GPT, Claude, and Deepseek without extra tokens or latency.