Comparing Bottom-Up and Top-Down Steering Approaches on In-Context Learning Tasks

Documenti analoghi

• Understanding (Un)Reliability of Steering Vectors in Language Models
  di: Braun, Joschka, et al.
  Pubblicazione: (2025)
• Implicit meta-learning may lead language models to trust more reliable sources
  di: Krasheninnikov, Dmitrii, et al.
  Pubblicazione: (2023)
• Fresh in memory: Training-order recency is linearly encoded in language model activations
  di: Krasheninnikov, Dmitrii, et al.
  Pubblicazione: (2025)
• Stress-Testing Capability Elicitation With Password-Locked Models
  di: Greenblatt, Ryan, et al.
  Pubblicazione: (2024)
• Hierarchical Synthetic Tabular Data Generation: A Hybrid Top-Down and Bottom-Up Framework
  di: Nie, Junfeng, et al.
  Pubblicazione: (2026)