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

    Neuronpedia

    INDEX
    Explanations

    help and advice

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

    phrases that signal formal, structured exposition—disclaimers, summaries, signposting, and instructional framing within a response.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
    AIRMAN
    0.54
    สต
    0.46
     الأخ
    0.45
     الاخ
    0.45
    BIN
    0.42
    ف
    0.41
    KU
    0.40
     INumber
    0.40
    ضيف
    0.39
     CORPER
    0.39
    POSITIVE LOGITS
     skilled
    0.48
     Village
    0.44
    食品
    0.44
     Detect
    0.43
     gauche
    0.43
     Crest
    0.43
     Skilled
    0.43
     Caball
    0.43
     Inspect
    0.43
     Vig
    0.42
    Act Density 0.022%

    No Known Activations