Position: Mechanistic Interpretability Should Prioritize Feature Consistency in SAEs
Paper
• 2505.20254 • Published
• 5
Zeyu Tang
zeyutang
·
AI & ML interests
Trustworthy AI
Recent Activity
new activity
2 days ago
aims-foundation/mac-fairness:Update 2-agent diffAware results (8 subcategories) new activity
2 days ago
aims-foundation/mac-fairness:Update 2-agent discrimEval results (2 subcategories) new activity
2 days ago
aims-foundation/mac-fairness:Update 2-agent BBQ results (11 subcategories) Organizations
authored a paper 9 months ago