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

    Neuronpedia

    INDEX
    Explanations

    phrases related to legal or formal matters

    oai_token-act-pair · gpt-3.5-turbo

    phrases that reference "regard" or "respect" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     tatt
    -0.70
     submar
    -0.67
     suspic
    -0.65
     helicop
    -0.65
     Monstrous
    -0.63
     Loaded
    -0.62
    avorite
    -0.62
     Sieg
    -0.62
     Panc
    -0.61
     hyster
    -0.61
    POSITIVE LOGITS
     thereto
    0.78
    ments
    0.78
     regards
    0.77
    rence
    0.76
    eous
    0.73
    erence
    0.73
    otaur
    0.71
     pertaining
    0.71
    forward
    0.70
    etics
    0.70
    Act Density 0.049%

    No Known Activations