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

    Neuronpedia

    INDEX
    Explanations

    Dates and updates

    np_max-act · gemini-2.0-flash

    The neuron activates on tokens that specify a temporal cutoff—particularly in “up to YYYY” date phrases indicating the model’s knowledge cutoff.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    ोत
    -0.07
     فت
    -0.07
    ptal
    -0.07
    igm
    -0.07
     노출등록
    -0.06
     summon
    -0.06
     BaseType
    -0.06
    systems
    -0.06
     addTarget
    -0.06
    EMENT
    -0.06
    POSITIVE LOGITS
     specifically
    0.07
     specializes
    0.07
     disple
    0.06
     esa
    0.06
    /my
    0.06
     experimented
    0.06
    üns
    0.06
     thảo
    0.06
    resenter
    0.06
    Nov
    0.06
    Act Density 0.014%

    No Known Activations