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

    Neuronpedia

    INDEX
    Explanations

    they say

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     dove
    0.60
    C
    0.58
     sot
    0.57
     maximum
    0.57
    业
    0.57
    พอ
    0.56
     glove
    0.56
    orde
    0.56
    প্রধানমন্ত্রী
    0.55
    이제
    0.55
    POSITIVE LOGITS
     say
    1.71
     says
    1.45
    say
    1.35
     Say
    1.23
     SAY
    1.20
    atically
    1.19
    說是
    1.17
    Say
    1.16
     saying
    1.15
    说是
    1.09
    Act Density 0.030%

    No Known Activations