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

    Neuronpedia

    INDEX
    Explanations

    personal names and decisions within informative texts

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating personal opinions or decisions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    .*
    -0.82
    .</
    -0.71
    manac
    -0.70
     .)
    -0.60
    .).
    -0.60
    .�
    -0.59
     ().
    -0.58
    *.
    -0.56
    âĢł
    -0.56
    Sample
    -0.56
    POSITIVE LOGITS
     warr
    0.65
     secondly
    0.64
     indemn
    0.63
     destro
    0.62
     alright
    0.61
     realise
    0.61
     apologise
    0.60
     hindsight
    0.60
     counselling
    0.59
    Secondly
    0.59
    Act Density 0.975%

    No Known Activations