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

    Neuronpedia

    INDEX
    Explanations

    terms or phrases that are persuasive, convincing, or provide strong evidence in different contexts

    oai_token-act-pair · gpt-3.5-turbo

    the word "compelling" and related concepts of persuasion or strong interest

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    hops
    -0.85
    pez
    -0.82
    hop
    -0.80
    atel
    -0.68
    pec
    -0.66
    ource
    -0.64
    ummer
    -0.63
    sterdam
    -0.63
    ipop
    -0.63
    chie
    -0.62
    POSITIVE LOGITS
    ly
    0.94
    ingly
    0.93
    enough
    0.91
    NESS
    0.82
    LY
    0.80
     enough
    0.77
    ively
    0.75
     compelling
    0.73
     persuasive
    0.72
    ments
    0.72
    Act Density 0.027%

    No Known Activations