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

    Neuronpedia

    INDEX
    Explanations

    phrases related to caution or prevention

    oai_token-act-pair · gpt-3.5-turbo

    negative constructions emphasizing avoidance or prohibition

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    BSD
    -0.71
    ciation
    -0.70
    Reviewer
    -0.69
     testament
    -0.67
    cyclopedia
    -0.67
    ledged
    -0.66
     doubtless
    -0.66
     Newsp
    -0.62
    upon
    -0.62
    eret
    -0.62
    POSITIVE LOGITS
     distractions
    0.84
     hesitate
    0.82
     reinvent
    0.81
     overly
    0.80
     necessarily
    0.80
     stray
    0.79
     interrupt
    0.79
     compromising
    0.77
     undue
    0.76
     rush
    0.75
    Act Density 0.379%

    No Known Activations