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

    Neuronpedia

    INDEX
    Explanations

    words that indicate a concession or opposing viewpoint

    oai_token-act-pair · gpt-3.5-turbo

    the word "though" indicating contrasts or exceptions in statements

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    asus
    -0.68
    urated
    -0.62
    tnc
    -0.61
    adesh
    -0.60
     Simulator
    -0.60
     GOODMAN
    -0.60
    enture
    -0.58
    otaur
    -0.58
    ragon
    -0.57
    enter
    -0.57
    POSITIVE LOGITS
    ts
    0.94
     admittedly
    0.86
    tons
    0.83
    abouts
    0.76
     thankfully
    0.74
     fortunately
    0.68
     ideally
    0.68
     interestingly
    0.67
    ^^^^
    0.66
     beware
    0.64
    Act Density 0.040%

    No Known Activations