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

    Neuronpedia

    INDEX
    Explanations

    numerical values related to quantities or comparisons. It looks for phrases like "More than," "According to," "Than," and "Less."

    oai_token-act-pair · gpt-3.5-turbo

    places in the text where formatting or special characters are used, likely focusing on paragraph breaks or sections

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    .","
    -0.71
     respectively
    -0.63
     prest
    -0.55
     thereto
    -0.53
     ..."
    -0.52
     âĢ¦"
    -0.52
     oxide
    -0.50
    ÏĢ
    -0.49
     conduc
    -0.49
    -"
    -0.49
    POSITIVE LOGITS
    resa
    1.00
    xiety
    0.85
    odore
    0.82
    notations
    0.70
    uckland
    0.70
     Reporter
    0.67
    anmar
    0.64
    mosp
    0.64
    bidden
    0.61
    iple
    0.61
    Act Density 0.533%

    No Known Activations