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

    Neuronpedia

    INDEX
    Explanations

    silly names and phrases

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     `
    0.40
     simplify
    0.38
     parameter
    0.38
    create
    0.38
     generate
    0.38
     create
    0.37
     use
    0.36
     Scalar
    0.36
    <code>
    0.36
     within
    0.36
    POSITIVE LOGITS
     shenanigans
    0.59
     extravaganza
    0.59
     antics
    0.56
     baddies
    0.55
     veggies
    0.53
     thugs
    0.52
     wacky
    0.50
     critters
    0.50
     puns
    0.49
     bunnies
    0.49
    Act Density 0.104%

    No Known Activations