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

    Neuronpedia

    INDEX
    Explanations

    comma or quote

    np_max-act · gemini-2.0-flash

    This neuron detects high‐level “meta” instructions or policy directives telling the assistant how to behave (e.g. “only send the completion based on the system instructions,” “don’t repeat,” etc.).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    .Bit
    -0.07
    .plugin
    -0.07
     виконав
    -0.07
     chips
    -0.07
    StackNavigator
    -0.06
     Falcon
    -0.06
     선수
    -0.06
    _column
    -0.06
    +j
    -0.06
    :uint
    -0.06
    POSITIVE LOGITS
     surv
    0.06
     Saskatchewan
    0.06
     Verify
    0.06
    .Find
    0.06
    ogi
    0.06
    'action
    0.06
     》
    0.06
     versatility
    0.06
     fade
    0.06
    gatsby
    0.06
    Act Density 0.026%

    No Known Activations