Apple、API 呼び出し型エージェント向け環境不要の合成データ生成法を発表

Training API-calling large language model (LLM) agents demands massive amounts of high-quality trajectories.

collecting such data at scale typically requires fully implemented environments with executable APIs and realistic, pre-populated backend databases

we propose an environment-free synthetic data generation approach that leverages LLMs as on-the-fly digital world models.