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

    Neuronpedia

    INDEX
    Explanations

    avoid unintended consequences

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    _(
    0.41
    --(
    0.40
    routes
    0.37
    nte
    0.37
    дый
    0.37
    },(
    0.36
    Notas
    0.36
    平台的
    0.36
    aturi
    0.35
     .(
    0.35
    POSITIVE LOGITS
     mischiev
    0.41
     mischief
    0.41
     കേ
    0.41
     Harlan
    0.40
     intervening
    0.40
     dut
    0.39
     jut
    0.39
     ausp
    0.39
     telescop
    0.39
     Undoubtedly
    0.39
    Act Density 0.000%

    No Known Activations