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

    Neuronpedia

    INDEX
    Explanations

    precise

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

    model-generated, structured technical output—especially code/markup blocks and chat/turn markers—rather than ordinary user prose.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
    icyclo
    0.44
    ambilan
    0.44
    荸
    0.43
     ост
    0.43
    ້ອງ
    0.41
    mirea
    0.41
     зага
    0.41
    :[/
    0.41
     вроде
    0.40
     امیدوار
    0.40
    POSITIVE LOGITS
    AUT
    0.46
    GRAY
    0.43
     REL
    0.41
    FUR
    0.41
     benc
    0.40
    LEV
    0.40
    XX
    0.39
     information
    0.39
     Esc
    0.39
     RES
    0.38
    Act Density 1.958%

    No Known Activations