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

    Neuronpedia

    INDEX
    Explanations

    phrases or sentences containing punctuation symbols followed by specific characters

    oai_token-act-pair · gpt-3.5-turbo

    instances of strong negative sentiments or failures

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     equip
    -0.68
     inverse
    -0.66
     lifes
    -0.64
     equival
    -0.63
    akedown
    -0.62
     legally
    -0.61
    boro
    -0.61
     alley
    -0.60
     associate
    -0.59
    uchs
    -0.59
    POSITIVE LOGITS
    Therefore
    0.85
    Nevertheless
    0.85
    However
    0.84
    Nonetheless
    0.83
    Firstly
    0.83
    ³³³³
    0.83
    SEE
    0.81
    Examples
    0.81
    Anonymous
    0.81
    RESULTS
    0.80
    Act Density 0.829%

    No Known Activations