Multi-domain reward models for chat responses. They use 23 fine-grained regression objectives and a prompt-conditioned gating network.