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

    Neuronpedia

    INDEX
    Explanations

    instances of attempted actions or efforts

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "tried."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    head
    -0.72
    scribe
    -0.71
    thus
    -0.70
    Production
    -0.69
    requisite
    -0.68
    performance
    -0.66
    Quality
    -0.64
     Instruction
    -0.63
    cised
    -0.62
    cedented
    -0.62
    POSITIVE LOGITS
     unsuccessfully
    1.24
     desperately
    0.81
     valiant
    0.79
     harder
    0.79
     tried
    0.78
    anke
    0.72
     repeatedly
    0.71
     prem
    0.67
    pload
    0.67
    secut
    0.65
    Act Density 0.044%

    No Known Activations