Direct-Inverse Prompting: Analyzing LLMs' Discriminative Capacity in Self-Improving Generation

Jihyun Janice Ahn, Ryo Kamoi, Lu Cheng +2

Mainstream LLM research has primarily focused on enhancing their generative capabilities. However, even the most advanced LLMs experience uncertainty in their outputs, often produc…