Distribution Prompting: Understanding the Expressivity of Language Models Through the Next-Token Distributions They Can Produce.

Haojin Wang, Zining Zhu, Freda Shi

Browse the full EMNLP paper archive.