Beyond Interpretability: When, Why, and How Sparse Autoencoders Enable Label-Free Visual Steering

Documenti analoghi

• Can Cross-Layer Transcoders Replace Vision Transformer Activations? An Interpretable Perspective on Vision
  di: Chatzoudis, Gerasimos, et al.
  Pubblicazione: (2026)
• LUCID-SAE: Learning Unified Vision-Language Sparse Codes for Interpretable Concept Discovery
  di: Gu, Difei, et al.
  Pubblicazione: (2026)
• The Hidden Life of Tokens: Reducing Hallucination of Large Vision-Language Models via Visual Information Steering
  di: Li, Zhuowei, et al.
  Pubblicazione: (2025)
• Test-Time Spectrum-Aware Latent Steering for Zero-Shot Generalization in Vision-Language Models
  di: Dafnis, Konstantinos M., et al.
  Pubblicazione: (2025)
• Training Like a Medical Resident: Context-Prior Learning Toward Universal Medical Image Segmentation
  di: Gao, Yunhe, et al.
  Pubblicazione: (2023)