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

    Neuronpedia

    INDEX
    Explanations

    soften

    np_max-act · gemini-2.0-flash

    The neuron activates on words that denote softening or rendering a material more pliable (e.g., “soften,” “pliable,” “softening”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    홍
    -0.06
    ortality
    -0.06
    osing
    -0.06
    目的
    -0.06
    �
    -0.06
     bevor
    -0.06
     nau
    -0.06
    Benef
    -0.06
     Bordeaux
    -0.06
     �
    -0.06
    POSITIVE LOGITS
    alarda
    0.07
    testdata
    0.07
     AT
    0.07
     Easily
    0.06
    ilver
    0.06
    {T
    0.06
    ianne
    0.06
    \""
    0.06
     kö
    0.06
    eses
    0.06
    Act Density 0.014%

    No Known Activations