4 ms·
Anecdata seems quite valid for LLM comparison when trying to evaluate 'usefullness' for users. The lmsys chat leaderboard is literally just mass anecdata.
by Chilko 2y ago
Anecdata seems quite valid for LLM comparison when trying to evaluate 'usefullness' for users. The lmsys chat leaderboard is literally just mass anecdata.
- lostmsu 2y agoYes, "mass anecdata" + blind collection is usually called "data".