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

    Neuronpedia

    INDEX
    Explanations

    Hmm, it appears that Neuron 4 is looking for patterns from a data source that may be encoded in a different format, possibly related to a specific language or system, as there doesn't seem to be a clear pattern based on the text provided

    oai_token-act-pair · gpt-3.5-turbo

    special characters or symbols with specific patterns

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    theless
    -0.71
    aturdays
    -0.70
    ographically
    -0.67
    ulators
    -0.67
     Kens
    -0.66
    ithing
    -0.64
     Ness
    -0.63
     Redd
    -0.63
    anwhile
    -0.61
    essa
    -0.60
    POSITIVE LOGITS
    ¹
    1.95
    ²
    1.88
    ¨
    1.86
    °
    1.85
    Ń
    1.84
    1.83
    ¦
    1.79
    ´
    1.76
    ¸
    1.76
    ¯
    1.76
    Act Density 0.007%

    No Known Activations