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

    Neuronpedia

    INDEX
    Explanations

    Deflection, diversion, or lies

    np_max-act · gemini-2.0-flash

    The neuron activates on instances of “from” used in contrast or diversion expressions (e.g. “different from,” “divert attention from,” “distract from the fact that”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    _Manager
    -0.07
    نا
    -0.07
    gons
    -0.06
    好像
    -0.06
    �
    -0.06
    iteral
    -0.06
     Israel
    -0.06
     Although
    -0.06
    ран
    -0.06
    rending
    -0.06
    POSITIVE LOGITS
     Adventure
    0.07
     prog
    0.07
     BANK
    0.06
     Accom
    0.06
     flight
    0.06
     Trip
    0.06
     HAPP
    0.06
    fcn
    0.06
    predictions
    0.06
     close
    0.06
    Act Density 0.017%

    No Known Activations