X：AK (@_akhaliq)· @HuggingPapers·· 2 天前AI 评分
新在线策略蒸馏方法：教师是方向而非终点，学生可匹配或超越教师
RT by @_akhaliq: The Teacher Is a Direction, Not a Destination New on-policy distillation method that extrapolates RL-induced representation residuals, letting students match or exceed the teacher across four model pairs. Code and checkpoints planned for release.
一种新的在线策略蒸馏方法通过外推 RL 引发的表征残差，让学生在四组模型对上匹配甚至超越教师模型。该方法将教师视为方向而非终点，代码和 checkpoint 计划发布。
来源：X：AK (@_akhaliq) · x.lingyaoai.com