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

    Neuronpedia

    INDEX
    Explanations

    this that helps makes works

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     when
    0.92
     during
    0.86
    During
    0.83
     lorsque
    0.81
     whenever
    0.79
     quando
    0.76
     যেসব
    0.75
    When
    0.75
     During
    0.74
    ),
    0.73
    POSITIVE LOGITS
     exceeds
    1.51
     satisfies
    1.51
     goes
    1.49
     gives
    1.45
     gets
    1.44
     contradicts
    1.44
     proves
    1.44
     constitutes
    1.43
     complies
    1.37
     makes
    1.37
    Act Density 0.075%

    No Known Activations