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

    Neuronpedia

    INDEX
    Explanations

    contrasting traditional with new

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    lisi
    0.37
    Param
    0.35
     behest
    0.35
     hatta
    0.35
     auch
    0.35
    paramet
    0.35
     multip
    0.34
    Regardless
    0.34
     bądź
    0.33
    bele
    0.33
    POSITIVE LOGITS
     Previously
    1.08
     previously
    1.03
    Previously
    0.98
     Whereas
    0.97
    previously
    0.96
    以前
    0.95
    従来の
    0.95
    従来
    0.93
     Unlike
    0.92
     traditionally
    0.92
    Act Density 0.275%

    No Known Activations