Toward faithful and human-aligned self-explanation of deep models – Nature
Toward faithful and human-aligned self-explanation of deep models – Nature