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

    Neuronpedia

    INDEX
    Explanations

    abstract states and comparisons

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    /
    0.53
     indicates
    0.50
    Additionally
    0.50
     btw
    0.50
    ak
    0.50
     FYI
    0.50
    有很多
    0.49
    ake
    0.49
    rm
    0.48
    am
    0.48
    POSITIVE LOGITS
     aquella
    0.64
     whispered
    0.61
    仿佛
    0.59
     настолько
    0.59
     whispering
    0.58
     aquel
    0.57
     마치
    0.57
     কিংবা
    0.56
     столь
    0.56
     मानो
    0.56
    Act Density 0.023%

    No Known Activations