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

    Neuronpedia

    INDEX
    Explanations

    listing explanations or options

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     truly
    1.07
     most
    1.02
     process
    1.01
     no
    1.00
     plus
    0.99
     concept
    0.99
     si
    0.98
     consider
    0.98
     due
    0.97
     best
    0.95
    POSITIVE LOGITS
    provide
    1.07
    have
    1.02
    Provide
    1.01
    that
    1.00
    Payments
    1.00
    using
    1.00
    Re
    0.99
    ideas
    0.98
    Using
    0.97
    the
    0.97
    Act Density 0.994%

    No Known Activations