Why do DeepLift, Guided BackProp and DeepLift produce similar explanation performances?

braindatalab / gecobench

NLP Benchmark for XAI methods

BSD 3-Clause "New" or "Revised" License

3 stars 1 forks source link

Closed rickwg closed 5 months ago

rickwg commented 9 months ago

Implementation of DeepLift, Guided BackProp and DeepLift checked. It is clear why they all produce the similar explanation performance.