1 paper · 1 filter
Berk Atil, Sarp Aykent, Alexa Chittams +10
LLM (large language model) practitioners commonly notice that outputs can vary for the same inputs under settings expected to be deterministic. Yet the questions of how pervasive t…