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

    Neuronpedia

    INDEX
    Explanations

    training was harsh

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     करतात
    1.46
     जातात
    1.45
     ചെയ്യുന്നു
    1.41
     помогают
    1.38
     sahiptir
    1.36
     येतात
    1.36
     imaju
    1.35
     होतात
    1.30
    ходят
    1.27
     असतात
    1.25
    POSITIVE LOGITS
     was
    4.84
    was
    3.80
     wasn
    3.75
     était
    3.44
     была
    3.38
     Was
    3.36
     был
    3.33
    Was
    3.32
     were
    3.28
     ήταν
    3.14
    Act Density 0.792%

    No Known Activations