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

    Neuronpedia

    INDEX
    Explanations

    new users

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

    code-related tokens, especially programming annotations/identifiers and large numeric literals within code blocks.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
     regs
    0.18
     multipurpose
    0.18
    VERTISING
    0.17
     $€
    0.17
     comparative
    0.17
     Oilers
    0.16
     statewide
    0.16
     grassroots
    0.16
     gigs
    0.16
     subsidies
    0.16
    POSITIVE LOGITS
    When
    0.24
    Because
    0.24
    In
    0.24
    While
    0.22
     Because
    0.21
    There
    0.20
    The
    0.19
    Note
    0.19
    However
    0.19
     In
    0.19
    Act Density 0.873%

    No Known Activations