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

    Neuronpedia

    INDEX
    Explanations

    contractions with 't, referring to negative actions or responses

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Negative Logits
    çĶŁ
    -0.82
    è¦ļéĨĴ
    -0.75
    å½
    -0.70
    Reviewer
    -0.69
    è»
    -0.66
    accompan
    -0.66
    ERG
    -0.66
     è£ıè
    -0.66
    stre
    -0.65
     Creat
    -0.65
    POSITIVE LOGITS
     necessarily
    1.30
     exactly
    1.10
     bother
    1.01
     quite
    0.97
     really
    0.93
     even
    0.92
     gotta
    0.91
     hesitate
    0.89
    icably
    0.89
    epad
    0.88
    Act Density 0.136%

    No Known Activations