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

    Neuronpedia

    INDEX
    Explanations

    who followed by verb

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    which
    0.67
     которая
    0.67
     который
    0.63
     которое
    0.63
     позволит
    0.63
     which
    0.60
     možete
    0.59
     която
    0.58
     budete
    0.58
    你會
    0.56
    POSITIVE LOGITS
     themselves
    1.06
     are
    0.98
     ovat
    0.90
     هستند
    0.89
     have
    0.86
     were
    0.84
     являются
    0.83
     aren
    0.83
     हैं
    0.80
     eivät
    0.79
    Act Density 0.100%

    No Known Activations