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

    Neuronpedia

    INDEX
    Explanations

    hit

    np_max-act · gemini-2.0-flash

    The neuron strongly activates on marketing-style endorsement language—i.e. first-person product sponsorship or promotional hype (words like “opportunity,” “have,” brand names, and praising verbs).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    pute
    -0.07
    .ErrorMessage
    -0.07
    ervlet
    -0.07
     insiders
    -0.06
    　r
    -0.06
    creature
    -0.06
     granny
    -0.06
    .filePath
    -0.06
    .protobuf
    -0.06
    рев
    -0.06
    POSITIVE LOGITS
    cerpt
    0.07
    .")↵
    0.07
     extended
    0.07
     korun
    0.06
    ("|
    0.06
     weiß
    0.06
     Auf
    0.06
    '))↵↵
    0.06
     momentum
    0.06
     dept
    0.06
    Act Density 0.100%

    No Known Activations