I looked at the supposed “research” article, there’s nothing to read except few charts to show off the “improvements” over current models. No discussion of the training method or dataset whatsoever.<p>If I remember correctly, the last decent research paper the company published was probably the InstructGPT paper.