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

    Neuronpedia

    INDEX
    Explanations

    ranking, superlative phrases

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     improper
    0.81
     incorrect
    0.77
     undesired
    0.76
     faulty
    0.71
     deleting
    0.70
     unspecified
    0.70
     decreases
    0.70
     unhealthy
    0.69
     erroneous
    0.68
     minor
    0.68
    POSITIVE LOGITS
     unparalleled
    1.47
     unrivalled
    1.44
     excellente
    1.43
     excellent
    1.41
     excelente
    1.38
     unrival
    1.38
     hervorrag
    1.37
     superbly
    1.37
    excellent
    1.35
     superb
    1.34
    Act Density 9.572%

    No Known Activations