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

    Neuronpedia

    INDEX
    Explanations

    tokenizer and tokenization

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    İ
    0.47
    ۱
    0.45
    P
    0.44
    Calc
    0.44
    S
    0.43
     =
    0.43
    見積
    0.43
     confirms
    0.42
     пребы
    0.42
    M
    0.41
    POSITIVE LOGITS
    0.49
     อีก
    0.47
    })}{\
    0.47
    囷
    0.45
    তাহাদের
    0.45
     এছাড়া
    0.45
     également
    0.44
    炆
    0.44
     देखील
    0.44
     flagging
    0.42
    Act Density 0.006%

    No Known Activations