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

    Neuronpedia

    INDEX
    Explanations

    instances where the word "probably" is used

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Negative Logits
    nan
    -1.11
    elight
    -1.08
    hips
    -1.07
    uctor
    -1.03
    issy
    -1.01
    lings
    -1.00
    vers
    -1.00
    ife
    -1.00
    arthed
    -1.00
    eem
    -0.99
    POSITIVE LOGITS
     underestimate
    1.05
    Ń·
    1.04
     regret
    1.03
     ali
    1.02
    ©¶æ
    0.99
     misunder
    0.97
     overest
    0.95
     quir
    0.95
     exagger
    0.94
     aval
    0.92
    Act Density 1.121%

    No Known Activations