XAI-Units offers synthetic datasets and handcrafted models with known ground-truth attributions to unit-test feature attribution methods, and demonstrates its use by revealing a normalization bug in Captum's DeepLIFT.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
XAI-Units: Benchmarking Explainability Methods with Unit Tests
XAI-Units offers synthetic datasets and handcrafted models with known ground-truth attributions to unit-test feature attribution methods, and demonstrates its use by revealing a normalization bug in Captum's DeepLIFT.