Reading Users' Minds from What They Say: An Investigation into LLM-based Empathic Mental Inference