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

    Neuronpedia

    INDEX
    Explanations

    Patent publications

    np_max-act · gemini-2.0-flash

    The neuron fires on numeric tokens, i.e. standalone numbers (years, patent/publication numbers, section or figure indices).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     LET
    -0.08
     seront
    -0.08
     tumblr
    -0.07
    Prof
    -0.07
    .elem
    -0.07
    .Selenium
    -0.07
     Synd
    -0.07
     Alone
    -0.06
     MOZ
    -0.06
     tavsiye
    -0.06
    POSITIVE LOGITS
    egal
    0.07
    icle
    0.06
    arah
    0.06
    τηκε
    0.05
    ))->
    0.05
    "`↵↵
    0.05
     prototype
    0.05
    itra
    0.05
     اختی
    0.05
    !");↵
    0.05
    Act Density 0.002%

    No Known Activations