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

    Neuronpedia

    INDEX
    Explanations

    sentences or phrases that end with periods, likely indicating the end of a complete thought or idea

    oai_token-act-pair · gpt-3.5-turbo

    sentences that mark the end of paragraphs or significant statements

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    adan
    -0.76
     candy
    -0.75
     invincible
    -0.74
     extinguished
    -0.74
     shake
    -0.70
     destroyer
    -0.69
     toast
    -0.69
     diminishing
    -0.69
     insur
    -0.68
     extinct
    -0.68
    POSITIVE LOGITS
     Specifically
    1.29
     Among
    1.24
     Essentially
    1.24
     Included
    1.22
     Basically
    1.18
     Readers
    1.16
    pdf
    1.16
     Apparently
    1.12
     Notably
    1.12
     Below
    1.10
    Act Density 0.433%

    No Known Activations