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

    Neuronpedia

    INDEX
    Explanations

    harm and negative consequences

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     letting
    0.81
    왤
    0.70
    太多
    0.69
     Too
    0.69
     Recent
    0.69
     too
    0.67
     Novel
    0.66
    Too
    0.65
     Uncommon
    0.65
    Recent
    0.65
    POSITIVE LOGITS
     ruined
    1.42
     decreased
    1.36
    reduced
    1.29
     reduced
    1.23
    increased
    1.18
     increased
    1.17
     erode
    1.12
    Reduced
    1.11
     destroyed
    1.11
     distorted
    1.10
    Act Density 0.311%

    No Known Activations