Jacob Hilton

reinforcement learning (3) reinforcement learning from human feedback (2) reward model (2) sample efficiency (2) model evaluation (1) falsehood detection (1) instruction following (1) language model alignment (1) model alignment (1) value function (1) off-policy learning (1) human feedback (1) language model (1) synthetic datum (1) scaling law (1) supervised fine-tuning (1) proximal policy optimization (1) on-policy learning (1) actor-critic method (1) feature distillation (1)