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

    Neuronpedia

    INDEX
    Explanations

    negations and situations with negative/potentially harmful connotations

    oai_token-act-pair · gpt-3.5-turbo

    negations and expressions of inability or lack

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    sis
    -0.72
     Invalid
    -0.70
     wording
    -0.69
    tein
    -0.68
    inning
    -0.65
     Reasons
    -0.65
    gar
    -0.63
    tons
    -0.63
     Cancel
    -0.62
    omen
    -0.62
    POSITIVE LOGITS
     necessarily
    0.99
     ever
    0.98
     bothered
    0.95
     bother
    0.95
     any
    0.94
     anyone
    0.93
     anything
    0.91
     anybody
    0.90
     overtly
    0.89
     remotely
    0.87
    Act Density 0.424%

    No Known Activations