Position: Mechanistic Interpretability Should Prioritize Feature Consistency in SAEs
Paper • 2505.20254 • Published • 5
Zeyu Tang
zeyutang
AI & ML interests
Trustworthy AI
Recent Activity
liked a dataset about 12 hours ago
actava/chi-bench upvoted a paper about 12 hours ago
CHI-Bench: Can AI Agents Automate End-to-End, Long-Horizon, Policy-Rich Healthcare Workflows? new activity 2 months ago
aims-foundations/mac-fairness:Update 1-agent summary (1 files)Organizations
authored a
paper 12 months ago