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

    Neuronpedia

    INDEX
    Explanations

    categorical

    np_max-act · gemini-2.0-flash

    The neuron activates on the token “categorical” (as in “categorical_crossentropy”), i.e. it detects that loss‐function keyword.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     sampler
    -0.06
     moons
    -0.06
    _PS
    -0.06
    .cell
    -0.06
     essence
    -0.06
     dessert
    -0.06
     outskirts
    -0.06
    (sound
    -0.06
    uses
    -0.06
    N
    -0.06
    POSITIVE LOGITS
     cmds
    0.07
    _mul
    0.07
     состоянии
    0.06
     süt
    0.06
     Terms
    0.06
     seria
    0.06
     saddened
    0.06
     Jag
    0.06
    怕
    0.06
    -navbar
    0.06
    Act Density 0.001%

    No Known Activations