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

    Neuronpedia

    INDEX
    Explanations

    to put it mildly

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     Use
    0.54
     use
    0.52
     using
    0.50
     enables
    0.49
     Using
    0.49
    Use
    0.49
     If
    0.48
     provides
    0.47
     improve
    0.44
     improves
    0.44
    POSITIVE LOGITS
     heartbreaking
    0.58
     shameful
    0.57
     horrors
    0.57
     भयानक
    0.57
     horrifying
    0.56
     absurdity
    0.53
     majest
    0.53
     horribly
    0.52
     terrifying
    0.52
     desolate
    0.52
    Act Density 0.003%

    No Known Activations