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

    Neuronpedia

    INDEX
    Explanations

    Website snippets from URLs

    np_max-act · gemini-2.0-flash

    The neuron activates on tokens that are parts of tech-news website URLs (e.g. “joystiq” or “engadget”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Raised
    -0.06
     every
    -0.06
     Paths
    -0.06
    _capability
    -0.06
     Benefits
    -0.06
     RTBU
    -0.06
    (term
    -0.06
     getSupportFragmentManager
    -0.06
     Rosie
    -0.06
     giờ
    -0.06
    POSITIVE LOGITS
     //~
    0.07
     сильно
    0.07
    .fil
    0.07
    ();//
    0.06
    .lr
    0.06
     //↵
    0.06
    _REUSE
    0.06
    -num
    0.06
     Precision
    0.06
     مجموعه
    0.06
    Act Density 0.059%

    No Known Activations