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

    Neuronpedia

    INDEX
    Explanations

    safe harbor, haven, zones

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     fatal
    0.44
    Fatal
    0.40
    faithful
    0.40
     Oft
    0.39
     Fatal
    0.39
    <unused63>
    0.39
     Save
    0.38
     Often
    0.38
    request
    0.38
    often
    0.38
    POSITIVE LOGITS
     harbor
    0.71
     haven
    0.66
     harbour
    0.63
     zones
    0.61
     harbors
    0.61
     Harbor
    0.60
    guards
    0.60
     harbours
    0.58
     Harbour
    0.57
    GUARD
    0.57
    Act Density 0.045%

    No Known Activations