Abstention Direction in Large Language Models: Mechanistic Interpretability of Abstention Across Query Ambiguity, Knowledge Limits, and Safety Alignment Paper • 2611.20441 • Published 4 days ago • 6
Mapping 1,000+ Language Models via the Log-Likelihood Vector Paper • 2502.16173 • Published Feb 22, 2025 • 2
Mapping 1,000+ Language Models via the Log-Likelihood Vector Paper • 2502.16173 • Published Feb 22, 2025 • 2