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

    Neuronpedia

    INDEX
    Explanations

    sequences of underscores and whitespace

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    Code-related terms and special characters

    np_acts-logits-general · gemini-2.0-flash

    C++ and Java syntax

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    ↵
    -0.23
     and
    -0.19
     –
    -0.18
    ↵↵
    -0.18
     Ch
    -0.17
     wasn
    -0.17
     isn
    -0.17
     see
    -0.17
    Ch
    -0.17
     ל
    -0.17
    POSITIVE LOGITS
    :✨
    1.21
     betweenstory
    1.00
     للاسماء
    0.94
    ᔢ
    0.93
     laſſen
    0.93
    <unused41>
    0.93
    <unused28>
    0.93
    <unused14>
    0.92
    <unused8>
    0.92
    [@BOS@]
    0.92
    Act Density 0.294%

    No Known Activations