A phase transition between positional and semantic learning in a solvable model of dot-product attention

Documenti analoghi

• Fundamental limits of learning in sequence multi-index models and deep attention networks: High-dimensional asymptotics and sharp thresholds
  di: Troiani, Emanuele, et al.
  Pubblicazione: (2025)
• Analysis of learning a flow-based generative model from limited sample complexity
  di: Cui, Hugo, et al.
  Pubblicazione: (2023)
• Dataset distillation for memorized data: Soft labels can leak held-out teacher knowledge
  di: Behrens, Freya, et al.
  Pubblicazione: (2025)
• Counting in Small Transformers: The Delicate Interplay between Attention and Feed-Forward Layers
  di: Behrens, Freya, et al.
  Pubblicazione: (2024)
• Quenches in the Sherrington-Kirkpatrick model
  di: Erba, Vittorio, et al.
  Pubblicazione: (2024)