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

    Neuronpedia

    INDEX
    Explanations

    Ignoring files or steps

    np_max-act · gemini-2.0-flash

    statements asserting that an action “can be” safely performed (e.g., removed/ignored/closed) without negative consequences.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
     intoxicated
    -0.06
     Included
    -0.06
     blooms
    -0.06
    .detectChanges
    -0.06
    -II
    -0.06
     Piano
    -0.06
    _sr
    -0.06
    ,R
    -0.06
    xmlns
    -0.06
     film
    -0.06
    POSITIVE LOGITS
    .getRoot
    0.08
    /arm
    0.07
    گار
    0.07
     thực
    0.06
    226
    0.06
     virtual
    0.06
    .can
    0.06
     конф
    0.06
    AUDIO
    0.06
     NDEBUG
    0.06
    Act Density 0.235%

    No Known Activations