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

    Neuronpedia

    INDEX
    Explanations

    statement or comment

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

    It detects long verbatim or block‑delimited quoted passages (text inside quotes/triple‑quotes or other context blocks).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Negative Logits
     músicas
    0.40
     geometries
    0.36
     cmds
    0.36
     échantillons
    0.36
     studs
    0.36
     regs
    0.36
     novas
    0.35
     pièces
    0.35
     cleats
    0.35
     infr
    0.35
    POSITIVE LOGITS
     statement
    0.45
     Comment
    0.39
    statement
    0.39
    这句话
    0.38
     Statement
    0.38
    comment
    0.38
     Paragraph
    0.37
     Description
    0.35
     Statements
    0.34
     Language
    0.34
    Act Density 0.717%

    No Known Activations