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

    Neuronpedia

    INDEX
    Explanations

    context-dependent effectiveness

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     handling
    0.43
     Handling
    0.42
     التعامل
    0.40
    灰
    0.39
     condimentum
    0.38
     руху
    0.38
    ponsorship
    0.37
    िख
    0.36
    Handling
    0.36
    羅
    0.36
    POSITIVE LOGITS
     hyvin
    0.67
     reliable
    0.64
     reliably
    0.63
     flawlessly
    0.61
     well
    0.60
     zuverläss
    0.59
    可靠
    0.57
     лучше
    0.56
     terbaik
    0.56
     bättre
    0.55
    Act Density 0.014%

    No Known Activations