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

    Neuronpedia

    INDEX
    Explanations

    software repositories

    np_max-act · gemini-2.0-flash

    The neuron is tuned to pick out pieces of URLs (protocol markers like “://”, domain fragments, file‐path segments, etc.).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     msgs
    -0.07
    經
    -0.06
     pauses
    -0.06
    .Print
    -0.06
     mastur
    -0.06
     Cut
    -0.06
    enas
    -0.06
    ared
    -0.06
    _Dis
    -0.06
     Prior
    -0.06
    POSITIVE LOGITS
    Criteria
    0.07
     Army
    0.06
    jsc
    0.06
    Miss
    0.06
    .lang
    0.06
     responders
    0.06
     helmets
    0.06
    PAL
    0.06
     cent
    0.06
    Lang
    0.06
    Act Density 0.006%

    No Known Activations