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

    Neuronpedia

    INDEX
    Explanations

    code

    np_max-act · gemini-2.0-flash

    The neuron responds to numeric literals in the text—i.e. tokens representing numbers (especially floating-point numbers).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     hóa
    -0.07
    desired
    -0.07
     SHOULD
    -0.07
    @@@@
    -0.07
     lamps
    -0.06
    	buff
    -0.06
     wires
    -0.06
     limitations
    -0.06
     nichž
    -0.06
    >false
    -0.06
    POSITIVE LOGITS
     Phaser
    0.06
     phoneNumber
    0.06
     erotik
    0.06
     Woman
    0.06
     Medi
    0.06
    Sel
    0.06
     القد
    0.06
     insane
    0.06
    -designed
    0.06
     Basketball
    0.06
    Act Density 0.046%

    No Known Activations