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

    Neuronpedia

    INDEX
    Explanations

    phrases or words indicating a sequence or order, specifically the word "Second" with a high activation value

    oai_token-act-pair · gpt-3.5-turbo

    phrases or words indicating a sequence or ranking

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     ga
    -0.79
    isconsin
    -0.65
    Sab
    -0.63
     pen
    -0.63
    renheit
    -0.60
    iesta
    -0.58
    nas
    -0.58
    IDA
    -0.57
     liter
    -0.56
    ihar
    -0.56
    POSITIVE LOGITS
     secondly
    0.75
     alternatively
    0.72
    appropriately
    0.72
     conclud
    0.69
     importantly
    0.68
     wors
    0.68
     worms
    0.64
     Artifact
    0.62
     cynicism
    0.62
     surpr
    0.61
    Act Density 0.115%

    No Known Activations