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

    Neuronpedia

    INDEX
    Explanations

    pronoun after dialogue punctuation

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     focuses
    0.53
    ："
    0.50
    ".
    0.50
     ->
    0.50
    :"
    0.50
     ".
    0.50
    '$
    0.49
    ...")
    0.48
     retains
    0.48
     --->
    0.48
    POSITIVE LOGITS
     he
    0.75
    他說
    0.75
    他说
    0.74
     she
    0.63
     murmured
    0.63
    she
    0.56
     он
    0.56
    她说
    0.54
     katanya
    0.53
     came
    0.49
    Act Density 0.009%

    No Known Activations