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

    Neuronpedia

    INDEX
    Explanations

    problems or flaws

    np_max-act · gemini-2.0-flash

    The neuron activates on critical and evaluative language highlighting weak or flawed evidence (e.g. terms like “rejecting,” “scientific evidence,” “negative consequences,” “harm,” “reasons why,” etc.).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    .getMethod
    -0.07
    CallableWrapper
    -0.06
    .cornerRadius
    -0.06
    Usuario
    -0.06
    belongsTo
    -0.06
    _hidden
    -0.06
     onPause
    -0.06
     ق
    -0.06
     tục
    -0.06
    .addTab
    -0.06
    POSITIVE LOGITS
    ϊ
    0.07
     vítěz
    0.07
    QUENCY
    0.07
     Chance
    0.07
    \">"
    0.07
    eways
    0.06
    RAFT
    0.06
    (fs
    0.06
     outsiders
    0.06
    <Real
    0.06
    Act Density 0.060%

    No Known Activations