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

    Neuronpedia

    INDEX
    Explanations

    how

    np_max-act · gemini-2.0-flash

    This neuron activates on parts of questions asking for instructions or processes, especially “how to” (and “where and how”) phrases.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    430
    -0.06
     Wyatt
    -0.06
    olly
    -0.06
    Subtitle
    -0.06
    Slide
    -0.06
    PasswordField
    -0.06
    ủng
    -0.06
     metric
    -0.06
     funcion
    -0.06
    em
    -0.06
    POSITIVE LOGITS
    =sub
    0.08
    �
    0.07
    keyup
    0.07
    /=
    0.07
    /$',
    0.06
     uid
    0.06
     robbed
    0.06
     retir
    0.06
    けて
    0.06
     transitioning
    0.06
    Act Density 0.005%

    No Known Activations