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

    Neuronpedia

    INDEX
    Explanations

    words related to expressing strong beliefs or support

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "espouse" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    graduate
    -0.75
    rooms
    -0.72
    Runner
    -0.70
    Reviewer
    -0.69
    ammy
    -0.66
    dry
    -0.65
    bread
    -0.65
    folk
    -0.64
     compuls
    -0.64
    fact
    -0.64
    POSITIVE LOGITS
     esp
    1.35
     mathemat
    1.19
    hovah
    1.01
    iscopal
    1.00
     advoc
    0.85
     challeng
    0.84
     tremend
    0.82
     acknow
    0.80
     preached
    0.79
    querque
    0.78
    Act Density 0.004%

    No Known Activations