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

    Neuronpedia

    INDEX
    Explanations

    quotation mark

    np_max-act · gemini-2.0-flash

    The neuron consistently activates on the JSON “author” field’s value, i.e. it detects when the text is the author entry in that metadata block.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    ومی
    -0.07
    fra
    -0.07
    AS
    -0.07
     майбут
    -0.06
    )/
    -0.06
    errer
    -0.06
     Khoa
    -0.06
    O
    -0.06
    ροι
    -0.06
    IAS
    -0.06
    POSITIVE LOGITS
    (enum
    0.07
     eax
    0.07
    산업
    0.06
    -readable
    0.06
    UPDATED
    0.06
     draw
    0.06
    /license
    0.06
    UpdateTime
    0.06
    [it
    0.06
     NSArray
    0.06
    Act Density 0.000%

    No Known Activations