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

    Neuronpedia

    INDEX
    Explanations

    input filtering and validation

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    safety
    0.46
     safety
    0.44
    Safety
    0.42
     dangereux
    0.42
    prone
    0.41
    dangerous
    0.40
    ിച്ചി
    0.40
     безопасность
    0.40
    SAFETY
    0.39
     susceptible
    0.39
    POSITIVE LOGITS
     purification
    0.73
     purifier
    0.68
     Purification
    0.67
     purifying
    0.65
     Sanit
    0.63
     FILTER
    0.61
     purify
    0.61
     Filter
    0.61
     filtration
    0.60
     filter
    0.59
    Act Density 0.004%

    No Known Activations