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

    Neuronpedia

    INDEX
    Explanations

    source code

    np_max-act · gemini-2.0-flash

    The neuron is keyed to numeric literals in code, i.e. any standalone number (integer or float) token.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    ア
    -0.07
    рукту
    -0.07
    598
    -0.06
    _dynamic
    -0.06
    xFA
    -0.06
    افته
    -0.06
    869
    -0.06
     Imaging
    -0.06
    ospital
    -0.06
     mannen
    -0.06
    POSITIVE LOGITS
     ){↵
    0.06
    rtle
    0.06
    =*/
    0.06
    =a
    0.06
     adapter
    0.06
    (left
    0.06
    _LAST
    0.06
    ..
    0.05
     locals
    0.05
     ){↵↵
    0.05
    Act Density 0.012%

    No Known Activations