1 paper
Kenan Hasanaliyev, Silas Alberti, Jenny Hamer +5
Reward models are a standard tool to score responses from LLMs. Reward models are built to rank responses to a fixed prompt sampled from a single model, for example to choose the b…