Showing cs.LGShow all
2 papers · 1 filter
cs.LG2026
The Optimal Decoder is Tilted Temperature
Nikolaos Papagiannis, Ananth Grama, Wojciech Szpankowski
Larger sample budgets benefit from higher sampling temperatures in LLMs, but what temperature tuning optimizes and why it works remain poorly understood. We seek to answer these qu…
cs.LG2024
No Free Lunch: Fundamental Limits of Learning Non-Hallucinating Generative Models
Changlong Wu, Ananth Grama, Wojciech Szpankowski
Generative models have shown impressive capabilities in synthesizing high-quality outputs across various domains. However, a persistent challenge is the occurrence of "hallucinatio…