the open
    research
    Contribute
    Back to the reading lists

    Safety, alignment & interpretability· 2018

    Sanity Checks for Saliency Maps

    Julius Adebayo (first author) · 2018

    Read the original

    arXiv 1810.03292OpenAlex W2891612330

    On the reading lists

    Why it matters

    Demonstrated that popular saliency methods can ignore model parameters, reshaping attribution evaluation.

    Condensed from the expert lists' notes.

    The record

    First released
    2018 (from arXiv 1810.03292)
    Citations
    2,947 citations, as counted by Semantic Scholar logoSemantic Scholar (non-commercial licence)
    Named by
    both expert lists
    Score
    0.620 · computed rank 130 · consensus 0.89, must-read 1, citations percentile 0.723 (2010s · safety), awards 0, existing lists 0. Method

    Editor's notes

    • Readability check: pass: A short, sobering test of explanation methods. (AI draft (Claude), not yet reviewed, 2026-10-29)

    Data sources

    Catalogue data from OpenAlex (CC0), Crossref (CC0 metadata; no abstracts are shown), arXiv, and Papers With Code (CC BY-SA 4.0), and Semantic Scholar logoSemantic Scholar (used under its non-commercial API licence). Each count names its source. See the method.