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

    Neuronpedia

    INDEX
    Explanations

    sum to 100

    np_max-act · gemini-2.0-flash

    This neuron detects phrasing around numeric totals summing to 100%, especially rounding‐related disclaimers.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     [...]↵↵
    -0.07
    okie
    -0.06
     gratitude
    -0.06
     Roe
    -0.06
     ト
    -0.06
     наличие
    -0.06
     Pole
    -0.06
     @
    -0.06
    ५
    -0.06
    @↵↵
    -0.06
    POSITIVE LOGITS
     violently
    0.07
    �
    0.07
     politik
    0.06
    gil
    0.06
     readFile
    0.06
    (gl
    0.06
     Carb
    0.06
     bağır
    0.06
     Turing
    0.06
    -court
    0.06
    Act Density 0.020%

    No Known Activations