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

    Neuronpedia

    INDEX
    Explanations

    can/cannot statements

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    Helping
    0.48
    promoting
    0.44
    HELP
    0.43
    Specific
    0.43
     chiar
    0.42
     TSS
    0.42
     자체가
    0.42
     Helping
    0.42
    你也
    0.41
    Chelsea
    0.41
    POSITIVE LOGITS
     however
    0.68
     však
    0.68
     όμως
    0.68
     പക്ഷേ
    0.61
     natomiast
    0.61
     gladly
    0.60
     però
    0.58
     azonban
    0.57
     viszont
    0.56
    却是
    0.56
    Act Density 0.003%

    No Known Activations