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

    Neuronpedia

    INDEX
    Explanations

    structural and formatting cues that organize text—like headings, bolded section titles, numbered lists, and table-style breakdowns.

    oai_token-act-pair · gpt-5Triggered by @johnathanlusun

    foundational, initial, basic

    np_max-act-logits · gemini-2.5-flashTriggered by @johnathanlusun
    New Auto-Interp
    Negative Logits
    Another
    0.64
    another
    0.60
     മറ്റൊരു
    0.60
     కూడా
    0.59
    Otro
    0.59
    other
    0.58
     другой
    0.57
    Otra
    0.57
    Otros
    0.57
    Other
    0.57
    POSITIVE LOGITS
     primary
    0.89
     beginner
    0.89
     både
    0.88
     zowel
    0.86
     both
    0.85
     basic
    0.83
     traditional
    0.83
     Traditional
    0.82
     classic
    0.82
     conventional
    0.82
    Act Density 3.939%

    No Known Activations