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

    Neuronpedia

    INDEX
    Explanations

    Japanese and Korean grammar particles

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     potentially
    0.53
    potentially
    0.47
     burgeoning
    0.47
     spelling
    0.46
     pesky
    0.46
     debilitating
    0.46
     groundbreaking
    0.45
     eponymous
    0.45
     doomed
    0.45
     escalating
    0.44
    POSITIVE LOGITS
    が
    0.85
    も
    0.80
    と
    0.80
    を
    0.78
    で
    0.78
    に
    0.75
    から
    0.73
    か
    0.73
    也
    0.73
    する
    0.71
    Act Density 0.015%

    No Known Activations