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

    Neuronpedia

    INDEX
    Explanations

    is followed by affirmation

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     horrible
    0.87
     stupid
    0.86
     crappy
    0.85
     messed
    0.84
     evil
    0.81
     crazy
    0.80
     generalizes
    0.77
     ruining
    0.76
     newbies
    0.74
     destroys
    0.74
    POSITIVE LOGITS
     remarkably
    1.35
     undoubtedly
    1.31
     undeniably
    1.30
     strikingly
    1.15
     unquestionably
    1.14
     decidedly
    1.09
     unmistak
    1.06
     markedly
    1.05
     eminently
    1.03
     testament
    1.03
    Act Density 0.617%

    No Known Activations