SAE Feature Interpretability
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
21.82Input Interpretability Score
9
Jun 9, 2026
6.96Interpretability Score
2
Jun 9, 2026
4.76Interpretability Score
2
Jun 9, 2026
5.3Interpretability Score
2
Jun 9, 2026
4.86Interpretability Score
2
Jun 9, 2026
5.44Interpretability Score
2
Jun 9, 2026
5.95Interpretability Score
2
Jun 9, 2026
4.02Interpretability Score
2
Jun 9, 2026