Mind the Gap: Conformative Decoding to Improve Output Diversity of Instruction-Tuned Large Language Models

Open in new window