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

    Neuronpedia

    INDEX
    Explanations

    categorized by how direct

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     когда
    0.43
     kenapa
    0.39
     cuando
    0.37
     why
    0.34
     eğer
    0.34
     когато
    0.34
     Why
    0.34
     quando
    0.34
    让她
    0.34
     اینکه
    0.33
    POSITIVE LOGITS
     they
    0.95
     we
    0.76
    they
    0.65
     each
    0.65
     it
    0.63
     они
    0.61
     he
    0.61
     THEY
    0.61
     you
    0.59
     вони
    0.59
    Act Density 0.042%

    No Known Activations