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

    Neuronpedia

    INDEX
    Explanations

    ambiguous words or phrases that can be open to interpretation

    oai_token-act-pair · gpt-3.5-turbo

    negations and phrases indicating what should or should not happen

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     Corona
    -0.70
     Wid
    -0.70
     satisfactory
    -0.63
     restores
    -0.61
     Niet
    -0.60
     clears
    -0.60
     restoration
    -0.59
     Fug
    -0.59
     enhancements
    -0.59
     Xuan
    -0.57
    POSITIVE LOGITS
     be
    0.80
     underestimate
    0.80
     bother
    0.76
     underestimated
    0.75
     blindly
    0.75
     interfere
    0.75
    erest
    0.74
    oga
    0.74
    reated
    0.74
     hesitate
    0.73
    Act Density 0.116%

    No Known Activations