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

    Neuronpedia

    INDEX
    Explanations

    AI refusal

    np_max-act-logits · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     Impacts
    0.59
     MPEG
    0.52
     Ans
    0.52
     Mitsubishi
    0.51
    铣
    0.50
     Mercedes
    0.50
     Helv
    0.50
     محم
    0.49
     एक्सपेरिमेंट
    0.49
     Epoxy
    0.49
    POSITIVE LOGITS
    RA
    0.77
     RA
    0.70
    rain
    0.57
    ra
    0.56
     Rail
    0.55
     ர
    0.55
    rail
    0.53
    Rail
    0.53
    RAT
    0.53
    ration
    0.52
    Act Density 0.147%

    No Known Activations