The tasks and counterfactuals from the Mechanistic Interpretability Benchmark.
Mechanistic Interpretability Benchmark
university
AI & ML interests
Principled evaluation of mechanistic interpretability methods.
Recent Activity
Collections
1
datasets
7
mib-bench/ravel
Viewer
•
Updated
•
132k
•
16
mib-bench/arc_easy
Viewer
•
Updated
•
4.01k
•
14
mib-bench/arc_challenge
Viewer
•
Updated
•
2k
•
10
mib-bench/copycolors_mcqa
Viewer
•
Updated
•
1.89k
•
32
mib-bench/arithmetic_subtraction
Viewer
•
Updated
•
22.4k
•
20
mib-bench/arithmetic_addition
Viewer
•
Updated
•
44.3k
•
14
mib-bench/ioi
Viewer
•
Updated
•
30k
•
86