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

    Neuronpedia

    INDEX
    Explanations

    Own

    np_max-act · gemini-2.0-flash

    The neuron spikes on numeric tokens expressing fractional or decimal values (i.e. numbers with a non‐zero decimal component).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    Cart
    -0.07
    Ex
    -0.07
     Wrong
    -0.07
    Ahead
    -0.07
    рест
    -0.07
     Goldberg
    -0.07
    mitted
    -0.06
    Adapter
    -0.06
    Prom
    -0.06
     Official
    -0.06
    POSITIVE LOGITS
     \<^
    0.07
     Lucas
    0.06
    asant
    0.06
    rawer
    0.06
    atie
    0.06
    ');↵↵↵↵
    0.06
    کان
    0.06
    stile
    0.06
    ↵↵↵↵
    0.06
     jun
    0.06
    Act Density 0.056%

    No Known Activations