Causal Attribution via Activation Patching

Documenti analoghi

• Understanding Counting Mechanisms in Large Language and Vision-Language Models
  di: Hasani, Hosein, et al.
  Pubblicazione: (2025)
• Uncovering Grounding IDs: How External Cues Shape Multimodal Binding
  di: Hasani, Hosein, et al.
  Pubblicazione: (2025)
• LibraGrad: Balancing Gradient Flow for Universally Better Vision Transformer Attributions
  di: Mehri, Faridoun, et al.
  Pubblicazione: (2024)
• Mechanistic Interpretability of Large-Scale Counting in LLMs through a System-2 Strategy
  di: Hasani, Hosein, et al.
  Pubblicazione: (2026)
• Visual Structures Helps Visual Reasoning: Addressing the Binding Problem in VLMs
  di: Izadi, Amirmohammad, et al.
  Pubblicazione: (2025)