MMODELYST
People/Qingyang Liu
PPL

Qingyang Liu

Papers (1)
The Missing Piece in Pre-trained Model Evaluation: Reward-Guided Decoding Unlocks Task-Oriented Behavior Without Parameter Updates