Archive
Fine-Tuning & Training

The Binary Loss Of Kahneman

Archive
Medium
150pts0 solves
Ethayarajh et al. (2024) proposed a preference-learning method that only needs binary thumbs-up/down labels (no paired chosen/rejected), inspired by a Nobel laureate's utility-theory ideas. Three-letter acronym. Flag format: CONGRESS{acronym}. Example: CONGRESS{dpo}.
Show hint
First letter is his surname's initial.

Archive — no submissions accepted

This challenge is preserved for reference. Play live challenges at /challenges.