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

    Neuronpedia

    INDEX
    Explanations

    Code Generation

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     assumes
    0.43
     goes
    0.38
     authorizes
    0.38
     gives
    0.36
     examines
    0.36
     использует
    0.36
    ohn
    0.35
     eliminates
    0.35
     Dynkin
    0.35
     deems
    0.34
    POSITIVE LOGITS
    的一种
    0.55
    ของการ
    0.49
    ของผู้
    0.47
     ενός
    0.45
    琰
    0.41
     හැකිය
    0.40
    。”
    0.39
    Prova
    0.39
     eines
    0.39
     μιας
    0.39
    Act Density 0.003%

    No Known Activations