Sam Marks comments on Recommend HAIST resources for assessing the value of RLHF-related alignment research

[Sam Marks]

Thanks, that’s a useful clarification; I’ll edit it into the post.