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

    Neuronpedia

    INDEX
    Explanations

    Snow

    np_max-act · gemini-2.0-flash

    The neuron strongly activates on numeral tokens (e.g. day/month/year numbers) – in other words, it picks out numeric date elements.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    marks
    -0.07
    Bid
    -0.07
     nova
    -0.07
     inadequate
    -0.07
     Fulton
    -0.06
     blueprint
    -0.06
    fection
    -0.06
    Œ
    -0.06
    íf
    -0.06
     neuronal
    -0.06
    POSITIVE LOGITS
     asymmetric
    0.08
    .visitVarInsn
    0.07
    ().'
    0.07
     disillusion
    0.07
    스가
    0.06
    ากล
    0.06
    .xhtml
    0.06
     반환
    0.06
    _cons
    0.06
    ());
↵
↵
    0.06
    Act Density 0.000%

    No Known Activations