Hey @sigchi ! I've only now read about the new "Assisted Desk Rejects" for @chi 2027 (via @tonofcrates ), and I _really_ think this process needs some more transparency: would you please make public a) the "rubric" or prompt, and b) which specific LLM you're using for the process?
2
Replies
0
Boosts
@floe @chi @tonofcrates There will be a response from the papers chairs soon!
@floe @sigchi @chi @tonofcrates
Hmm. I'd guess that making this information public would also make it easier for authors to engineer their paper so that it passes the LLM evaluator.
Thus, the human reviewers will get to read papers optimized for LLM classification. Not sure whether this means that papers will get better or worse through this process.
