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

    Neuronpedia

    INDEX
    Explanations

    strongly recommend

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     becomes
    1.19
     was
    1.16
     is
    1.15
    was
    1.07
    becomes
    1.05
     становится
    1.04
     էր
    1.02
     wasnt
    1.01
     isn
    1.00
     է
    0.99
    POSITIVE LOGITS
     intend
    1.80
     have
    1.79
     believe
    1.78
     owe
    1.71
     want
    1.62
     anticipate
    1.56
     expect
    1.54
     perceive
    1.51
    have
    1.50
     rely
    1.50
    Act Density 0.343%

    No Known Activations