Transformer learns the cross-task prior and regularization for in-context learning

Documenti analoghi

• Breaking through the learning plateaus of in-context learning in Transformer
  di: Fu, Jingwen, et al.
  Pubblicazione: (2023)
• Optimal cross-learning for contextual bandits with unknown context distributions
  di: Schneider, Jon, et al.
  Pubblicazione: (2024)
• On the importance of cross-task features for class-incremental learning
  di: Soutif--Cormerais, Albin, et al.
  Pubblicazione: (2021)
• Hi-fi functional priors by learning activations
  di: Sendera, Marcin, et al.
  Pubblicazione: (2025)
• On learning functions over biological sequence space: relating Gaussian process priors, regularization, and gauge fixing
  di: Petti, Samantha, et al.
  Pubblicazione: (2025)