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

    Neuronpedia

    INDEX
    Explanations

    punctuation marks such as periods and question marks

    oai_token-act-pair · gpt-3.5-turbo

    sentences that indicate continuation or conclusion

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     landfall
    -0.73
    ricanes
    -0.72
    ¥ŀ
    -0.71
     prohib
    -0.70
     disemb
    -0.70
     deliber
    -0.70
     âĢ
    -0.68
     quir
    -0.68
     responders
    -0.68
     emanc
    -0.67
    POSITIVE LOGITS
     Anyway
    1.37
     Lastly
    1.24
    Anyway
    1.23
     Secondly
    1.23
     Basically
    1.13
     Eventually
    1.13
     Needless
    1.12
     Finally
    1.12
     Then
    1.10
     Obviously
    1.10
    Act Density 0.554%

    No Known Activations