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

    Neuronpedia

    INDEX
    Explanations

    Surface

    np_max-act · gemini-2.0-flash

    This neuron activates on numerical tokens (digits or numbers) in math-problem text.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Harold
    -0.08
     <->
    -0.08
     dram
    -0.06
     Floating
    -0.06
    Init
    -0.06
     signal
    -0.06
     Drama
    -0.06
    สำค
    -0.06
    Traversal
    -0.06
     Woj
    -0.06
    POSITIVE LOGITS
    ’une
    0.07
    パン
    0.07
    π
    0.06
    IFDEF
    0.06
    /controllers
    0.06
    曜日
    0.06
    rientation
    0.06
     ↓
    0.06
     wollen
    0.06
    _NUMBER
    0.06
    Act Density 0.016%

    No Known Activations