the open
    research
    Contribute
    Back to the reading lists

    Safety, alignment & interpretability· 2022

    Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small

    Kevin Wang (first author) · 2022

    Read the original

    arXiv 2211.00593OpenAlex W4308023630

    On the reading lists

    Why it matters

    Reverse-engineered a concrete language-model behavior using attention heads and causal interventions.

    Condensed from the expert lists' notes.

    The record

    First released
    2022 (from arXiv 2211.00593)
    Citations
    1,377 citations, as counted by Papers With Code (CC BY-SA 4.0)
    Named by
    the chatgpt list only
    Score
    0.313 · computed rank 575 · consensus 0.3, must-read 0, citations percentile 0.638 (2020s · safety), awards 0, existing lists 0. Method

    Data sources

    Catalogue data from OpenAlex (CC0), Crossref (CC0 metadata; no abstracts are shown), arXiv, and Papers With Code (CC BY-SA 4.0). Each count names its source. See the method.