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

    Neuronpedia

    INDEX
    Explanations

    phrases indicating a requirement or condition that needs to be fulfilled before something else can happen

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Negative Logits
    ãĥĺ
    -0.75
    ãģ®æ
    -0.73
    ardless
    -0.72
    ãģ®éŃĶ
    -0.72
    raq
    -0.70
    rolled
    -0.67
    geist
    -0.67
    eller
    -0.67
    iph
    -0.66
    atform
    -0.64
    POSITIVE LOGITS
     risking
    1.20
     encountering
    1.04
     mentioning
    1.00
     sacrificing
    1.00
     recourse
    0.98
     consulting
    0.94
     knowing
    0.93
     compromising
    0.93
     violating
    0.90
     addressing
    0.89
    Act Density 0.059%

    No Known Activations