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

    Neuronpedia

    INDEX
    Explanations

    keep momentum and state

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     geworden
    0.59
     Expected
    0.48
     पहला
    0.44
    ที่มี
    0.44
    始まり
    0.44
    强化
    0.43
     예상
    0.43
    expected
    0.42
    新型
    0.42
     নির্ধার
    0.42
    POSITIVE LOGITS
     keep
    0.86
     afloat
    0.81
    keep
    0.80
    保持
    0.79
     maintain
    0.75
    Maintain
    0.75
    Keep
    0.74
     maintains
    0.73
     Keep
    0.71
     intact
    0.70
    Act Density 0.044%

    No Known Activations