AI brief
The author introduced Carrot-Parsnip, a simple social deduction game, and reports that most models perform better than random at identifying the hidden deceptive role (Parsnips) and that some are skilled at strategic deception.
Written by AI from LessWrong's published text.
Source
“TL:DR Earlier this year I introduced the very simple Social Deduction game of Carrot-Parsnip and showed that most models are better-than-random at identifying Parsnips (the hidden, deceptive role) and that being a strong Parsnip (skilled at strategic…”
Interpretation or community commentary rather than straight reporting.
Read original story ↗