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

    Neuronpedia

    INDEX
    Explanations

    vanishing gradient problems

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    IK
    0.46
    カロ
    0.45
    styled
    0.45
    密
    0.44
    宓
    0.44
    พิมพ์
    0.43
    PY
    0.42
    SK
    0.42
    ℍ
    0.41
    数字
    0.40
    POSITIVE LOGITS
     wacht
    0.53
     quando
    0.48
     esperando
    0.48
     cuando
    0.47
     backtracking
    0.47
     waiting
    0.47
     prilikom
    0.46
     Cuando
    0.46
     relembrar
    0.45
     evitare
    0.44
    Act Density 0.156%

    No Known Activations