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

    Neuronpedia

    INDEX
    Explanations

    G

    np_max-act · gemini-2.0-flash

    This neuron responds to structural or metadata markers (e.g. system/user/assistant tags and special boundary tokens) rather than the actual text content.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     abbreviation
    -0.07
     Ages
    -0.07
     Mona
    -0.07
     anymore
    -0.07
    ,"%
    -0.06
    ifes
    -0.06
     Alley
    -0.06
     recovered
    -0.06
     pione
    -0.06
    eldo
    -0.06
    POSITIVE LOGITS
    .DateField
    0.06
     endforeach
    0.06
    0.06
     selon
    0.06
    Https
    0.06
    едини
    0.06
     RESP
    0.06
     후
    0.06
     tienen
    0.06
     helm
    0.06
    Act Density 0.098%

    No Known Activations