Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    INDEX
    Explanations

    Code/data delimiters

    np_max-act · gemini-2.0-flash

    The neuron fires on quoted words used as keys in JSON‐style key/value pairs (e.g. property names like "name", "email", "subject", "predicate", "object" immediately before a colon).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    ruk
    -0.07
    685
    -0.07
    quer
    -0.07
     taxi
    -0.07
    ipp
    -0.07
    績
    -0.07
     Window
    -0.07
     Crate
    -0.07
    037
    -0.07
    bbbb
    -0.07
    POSITIVE LOGITS
     zpráv
    0.06
     SHARES
    0.06
    ----↵↵
    0.06
     соответств
    0.06
     Redskins
    0.06
    Gtk
    0.05
     dumps
    0.05
     Protect
    0.05
    0.05
    _InternalArray
    0.05
    Act Density 0.052%

    No Known Activations