LightningRodLabs/future-as-label-paper-step160 Reinforcement Learning • 33B • Updated Jan 16 • 18 • 3
view reply isn't "in context learning rate" is about training dynamics rather than the model architecture itself?