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

    Neuronpedia

    INDEX
    Explanations

    Non-English text

    np_max-act · gemini-2.0-flash

    The neuron is keyed to the Unicode replacement character (�) and other out‐of‐vocabulary or garbled tokens, effectively flagging decoding errors or unrecognized characters.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    IDS
    -0.07
    ()[
    -0.07
    Foo
    -0.06
     Hire
    -0.06
     payloads
    -0.06
    pause
    -0.06
    分布
    -0.06
    ída
    -0.06
    Oil
    -0.06
    ells
    -0.06
    POSITIVE LOGITS
    QUERY
    0.07
    xcb
    0.06
     coordinate
    0.06
    OWN
    0.06
     honeymoon
    0.06
     undermin
    0.06
     forall
    0.06
    own
    0.06
    (register
    0.06
    VIN
    0.06
    Act Density 0.005%

    No Known Activations