half-kh-hacker·قبل 4 أشهر·discussHow does post-training via reinforcement learning factor in? Does every evaluated judgement count as 'the training data' ?
half-kh-hacker·قبل 9 أشهر·discussthe paper already says "Butter-Bench evaluates a model's ability to 'pass the butter' (Adult Swim, 2014)" so