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

    Neuronpedia

    INDEX
    Explanations

    think what’s happening

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     cloche
    -1.48
     chemise
    -1.47
    MODO
    -1.38
     Requis
    -1.36
    げる
    -1.34
     Letra
    -1.32
    きましたが
    -1.31
     Obt
    -1.30
    िक्ष
    -1.29
    今では
    -1.27
    POSITIVE LOGITS
     for
    1.62
     what
    1.60
     —
    1.47
     and
    1.44
     –
    1.42
     its
    1.35
     [
    1.33
    ldots
    1.33
     s
    1.30
     R
    1.30
    Act Density 0.014%

    No Known Activations