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

    Neuronpedia

    INDEX
    Explanations

    describing explaining reading

    np_max-act · gemini-2.0-flash

    The neuron strongly activates on the occurrence of the tokens “description” or “descriptions,” i.e. when the text is issuing instructions to describe or give descriptions.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     NSK
    -0.07
    reatment
    -0.06
    ्यव
    -0.06
     vybav
    -0.06
     commanded
    -0.06
     Commander
    -0.06
    这样
    -0.06
     blowjob
    -0.06
     whip
    -0.06
    CARD
    -0.06
    POSITIVE LOGITS
    <script
    0.07
    \Client
    0.07
     "/
    0.07
    (min
    0.07
    CFG
    0.06
     tub
    0.06
     Gong
    0.06
    /train
    0.06
    akin
    0.06
    zhou
    0.06
    Act Density 0.010%

    No Known Activations