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

    Neuronpedia

    INDEX
    Explanations

    phrases related to completing a task or reaching a goal

    oai_token-act-pair · gpt-3.5-turbo

    instances of achievement or completion

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    ials
    -0.64
     Slate
    -0.63
     fluct
    -0.62
    Spread
    -0.62
     constantly
    -0.61
     Rue
    -0.60
    conom
    -0.58
    flat
    -0.58
     Scrib
    -0.58
     alone
    -0.56
    POSITIVE LOGITS
     relent
    0.75
     resil
    0.74
    atown
    0.71
    onso
    0.70
    iflower
    0.69
    aterasu
    0.68
     reckoning
    0.68
    ppo
    0.67
    vana
    0.66
    ileaks
    0.65
    Act Density 0.155%

    No Known Activations