search

dilyabareeva / quanda

A toolkit for quantitative evaluation of data attribution methods.
https://quanda.readthedocs.io
MIT License
33 stars 0 forks source link

Explanation Dataset should not be a part of Benchmark.evaluate #129

Closed dilyabareeva closed 2 months ago

dilyabareeva commented 2 months ago

We only want to evaluate a specific explainer, not a test dataset.