A Unified Theory of Sparse Dictionary Learning in Mechanistic Interpretability: Piecewise Biconvexity and Spurious Minima

Documenti analoghi

• Bridging Mechanistic Interpretability and Prompt Engineering with Gradient Ascent for Interpretable Persona Control
  di: Saini, Harshvardhan, et al.
  Pubblicazione: (2026)
• When Language Overwrites Vision: Over-Alignment and Geometric Debiasing in Vision-Language Models
  di: Saini, Harshvardhan, et al.
  Pubblicazione: (2026)
• AdaptiveK: Complexity-Driven Sparse Autoencoders for Interpretable Language Model Representations
  di: Yao, Yifei, et al.
  Pubblicazione: (2025)
• An Augmentation-Aware Theory for Self-Supervised Contrastive Learning
  di: Cui, Jingyi, et al.
  Pubblicazione: (2025)
• Biconvex Biclustering
  di: Rosen, Sam, et al.
  Pubblicazione: (2026)