Penalty Decoding: Well Suppress the Self-Reinforcement Effect in Open-Ended Text Generation

Open in new window