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

    Neuronpedia

    INDEX
    Explanations

    Determining start and end points

    np_max-act · gemini-2.0-flash

    This neuron is most responsive to interrogative and relative words—e.g. “what,” “which,” “when,” “how,” “ends,” “starts”—that introduce questions or clause‐boundary descriptions.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    .Ticks
    -0.07
     ва
    -0.06
     nuova
    -0.06
    刚才
    -0.06
    solver
    -0.06
     beh
    -0.06
     kter
    -0.06
     بايد
    -0.06
    工业
    -0.06
    кові
    -0.06
    POSITIVE LOGITS
    setCurrent
    0.06
     запах
    0.06
    恒
    0.06
     Nexus
    0.06
    kont
    0.06
    .'/
    0.06
    (nd
    0.06
     Murdoch
    0.06
     eventos
    0.06
    _frac
    0.06
    Act Density 0.151%

    No Known Activations