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

    Neuronpedia

    INDEX
    Explanations

    issues and claims

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    そして
    0.50
    และ
    0.50
     Universe
    0.48
    です
    0.46
     svojim
    0.44
    whose
    0.44
    。
    0.44
     Most
    0.43
    。“
    0.43
     Because
    0.42
    POSITIVE LOGITS
    <unused2157>
    0.52
     related
    0.50
     toward
    0.49
     akin
    0.49
     involving
    0.49
    abilité
    0.48
     relating
    0.47
     통한
    0.47
     pertaining
    0.46
    cribing
    0.45
    Act Density 0.395%

    No Known Activations