TAPR: A New Model That Rewrites Prompts to Boost LLM Performance

TAPR is an AI model that rewrites user prompts into clearer, more effective versions, aiming to improve the performance of Large Language Models (LLMs) on various tasks.

TAPR is trained using reinforcement learning (GRPO). It is rewarded when its rewritten prompts lead to better outcomes from an LLM, with another LLM acting as a judge to evaluate the quality.

TAPR helps companies use LLMs more efficiently without needing prompt engineering experts, reducing costs and speeding up the development of applications in areas like customer service and data analysis.

Yes, the code is publicly available on GitHub (https://github.com/OliverSavolainen/task-specific-prompt-rewriter), allowing researchers and developers to test and build upon it.