Transformers Provably Learn Algorithmic Solutions for Graph Connectivity, But Only with the Right Data

Documenti analoghi

• Transformers Learn to Achieve Second-Order Convergence Rates for In-Context Linear Regression
  di: Fu, Deqing, et al.
  Pubblicazione: (2023)
• Pre-trained Large Language Models Use Fourier Features to Compute Addition
  di: Zhou, Tianyi, et al.
  Pubblicazione: (2024)
• Convergent Evolution: How Different Language Models Learn Similar Number Representations
  di: Fu, Deqing, et al.
  Pubblicazione: (2026)
• FoNE: Precise Single-Token Number Embeddings via Fourier Features
  di: Zhou, Tianyi, et al.
  Pubblicazione: (2025)
• Transformers Learn Low Sensitivity Functions: Investigations and Implications
  di: Vasudeva, Bhavya, et al.
  Pubblicazione: (2024)