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

    Neuronpedia

    INDEX
    Explanations

    words related to discussion, debate, or argument

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to moral or ethical implications

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     recovered
    -0.76
    stood
    -0.76
    effects
    -0.71
     survives
    -0.70
     stabilized
    -0.65
     reusable
    -0.65
     survived
    -0.64
    ukong
    -0.64
     retained
    -0.64
    »Ĵ
    -0.63
    POSITIVE LOGITS
     folly
    1.21
     disingen
    1.20
     foolish
    1.15
     blasphemy
    1.15
     irresponsible
    1.13
     heresy
    1.12
    udicrous
    1.11
     delusional
    1.10
     dishon
    1.10
     insulting
    1.08
    Act Density 0.236%

    No Known Activations