This is definitely a promising next direction. One lesson from working on the diff between chat and base is that the difference is not ‘localized’ enough: chat and base have too many differences. Taking checkpoints that are closer together can improve on this.
This is definitely a promising next direction. One lesson from working on the diff between chat and base is that the difference is not ‘localized’ enough: chat and base have too many differences. Taking checkpoints that are closer together can improve on this.