ExpertLens: Activation steering features are highly interpretable

Documenti analoghi

• Steering into New Embedding Spaces: Analyzing Cross-Lingual Alignment Induced by Model Interventions in Multilingual Language Models
  di: Sundar, Anirudh, et al.
  Pubblicazione: (2025)
• What do your logits know? (The answer may surprise you!)
  di: Fedzechkina, Masha, et al.
  Pubblicazione: (2026)
• LinEAS: End-to-end Learning of Activation Steering with a Distributional Loss
  di: Rodriguez, Pau, et al.
  Pubblicazione: (2025)
• Investigating Intersectional Bias in Large Language Models using Confidence Disparities in Coreference Resolution
  di: Khan, Falaah Arif, et al.
  Pubblicazione: (2025)
• Discriminating Form and Meaning in Multilingual Models with Minimal-Pair ABX Tasks
  di: de Seyssel, Maureen, et al.
  Pubblicazione: (2025)