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

    Neuronpedia

    INDEX
    Explanations

    keep verb + gerund/preposition

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    Stay
    0.61
     remains
    0.54
     stays
    0.53
     stay
    0.52
    保持
    0.50
    remain
    0.50
     Stay
    0.48
    Keep
    0.48
    Remain
    0.47
     Remain
    0.46
    POSITIVE LOGITS
     FreeBSD
    0.44
    失败
    0.44
     postponing
    0.43
    ီ
    0.42
     purposely
    0.39
    失敗
    0.39
     zeggen
    0.39
    ﺰ
    0.39
    捅
    0.38
     fail
    0.38
    Act Density 0.011%

    No Known Activations