AI 导读
Trajectory 在雄心勃勃的目标上展现出的品味与执行力，总体上让我印象深刻。在我们的持续学习（Continual Learning）专场中，@rronak_ 非常用心地概述了他们如何应对 CL 中遗留的主要数据问题，包括为什么 GRPO 不够用、他们不得不转向 on-policy……然后再逐一解决由此产生的所有问题 这是来自该领域早期领军者之一的不错概述！ （更多内容见该专场其余部分，这一场阵容相当强）
正文
Trajectory have generally impressed me with their tasteful execution on ambitious goals. on our Continual Learning track, @rronak_ gave a very thoughtful overview on how they're tackling the main data problems left in CL, including why GRPO isn't enough and they had to go on-policy.... and then subsequently fix all the issues that come up with it
nice overview from one of the early leaders in this field!
(see the rest of the track for more, this one was quite stacked)
来源：@swyx · x.com