ARM: Efficient Guided Decoding with Autoregressive Reward Models

Open in new window