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

    Neuronpedia

    INDEX
    Explanations

    enumerating aspects

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     помочь
    0.49
     помога
    0.44
     realmente
    0.43
     really
    0.43
     khiến
    0.43
     truly
    0.41
     actively
    0.40
     réellement
    0.40
     aiutare
    0.40
     genuinely
    0.40
    POSITIVE LOGITS
    不但
    1.12
    不仅
    1.02
    不僅
    0.88
    一方面
    0.79
     nejen
    0.75
     একদিকে
    0.66
    既
    0.66
    一是
    0.58
     både
    0.57
     zowel
    0.57
    Act Density 0.048%

    No Known Activations