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

    Neuronpedia

    INDEX
    Explanations

    variations or distinctions in behavior or treatment among subjects

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    contrasting how things are done

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     =
    -0.48
    ⇒
    -0.42
    an
    -0.41
    *
    -0.38
    chließ
    -0.37
    com
    -0.37
    =
    -0.37
     of
    -0.36
    0
    -0.36
    िल्म
    -0.36
    POSITIVE LOGITS
     differently
    1.80
     anders
    1.01
     differentially
    0.98
     Differ
    0.94
     diffé
    0.93
     inaczej
    0.91
    differ
    0.91
    Differ
    0.79
     autrement
    0.78
     DIFFERENT
    0.73
    Act Density 0.006%

    No Known Activations