Researchers tested using large language models to generate planning domain models from text descriptions, and found larger models showed promise. via Hugging Face.