A circuit for predicting hierarchical structure in-context in Large Language Models

Documenti analoghi

• Sparse Autoencoders Reveal Temporal Difference Learning in Large Language Models
  di: Demircan, Can, et al.
  Pubblicazione: (2024)
• Simplifying Latent Dynamics with Softly State-Invariant World Models
  di: Saanum, Tankred, et al.
  Pubblicazione: (2024)
• Next state prediction gives rise to entangled, yet compositional representations of objects
  di: Saanum, Tankred, et al.
  Pubblicazione: (2024)
• Evaluating alignment between humans and neural network representations in image-based learning tasks
  di: Demircan, Can, et al.
  Pubblicazione: (2023)
• Reference-Free Rating of LLM Responses via Latent Information
  di: Girrbach, Leander, et al.
  Pubblicazione: (2025)