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

    Neuronpedia

    INDEX
    Explanations

    starting quotes

    np_acts-logits-general · gemini-2.5-flash-lite

    opening quotation marks indicating the start of quoted text or string literals, especially in code snippets.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
     harmonics
    0.26
     wells
    0.25
     savory
    0.25
     conund
    0.23
     pods
    0.23
     buds
    0.22
     savvy
    0.22
     analogs
    0.22
     pitfalls
    0.22
     savoury
    0.22
    POSITIVE LOGITS
    A
    0.39
    The
    0.38
    I
    0.38
    Please
    0.34
    T
    0.34
    This
    0.32
    What
    0.32
    There
    0.31
    How
    0.30
    Two
    0.29
    Act Density 0.199%

    No Known Activations