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

    Neuronpedia

    INDEX
    Explanations

    phrases or sentences starting with "Well" and involving a dialogue or conversation

    oai_token-act-pair · gpt-3.5-turbo

    conversational elements, particularly responses that begin with "Well" and other introductory phrases

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     clut
    -0.61
    @@
    -0.61
     buggy
    -0.57
     sway
    -0.57
     Sed
    -0.56
     shroud
    -0.55
    Âł
    -0.55
     tab
    -0.54
     BC
    -0.54
     polluted
    -0.54
    POSITIVE LOGITS
    resents
    0.71
    resa
    0.70
    zb
    0.70
    glas
    0.69
    ttes
    0.69
    resy
    0.68
    zbollah
    0.66
    eworld
    0.66
    iago
    0.66
    ocument
    0.65
    Act Density 0.182%

    No Known Activations