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

    Neuronpedia

    INDEX
    Explanations

    contractions with 's (is or has)

    oai_token-act-pair · gpt-3.5-turbo

    repetitive phrases that instruct or initiate actions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    ELD
    -0.85
    eur
    -0.77
    lessly
    -0.67
     biod
    -0.62
    /
    -0.62
     surrounds
    -0.61
    rupted
    -0.60
    oft
    -0.60
     haun
    -0.60
    Creat
    -0.57
    POSITIVE LOGITS
     suppose
    1.01
     assume
    0.98
     pretend
    0.98
     discuss
    0.88
     speculate
    0.86
     presume
    0.86
     gotta
    0.84
     get
    0.81
     go
    0.79
     celebrate
    0.79
    Act Density 0.020%

    No Known Activations