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

    Neuronpedia

    INDEX
    Explanations

    question prompts

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

    section headers and metadata that mark the start of a question prompt in chat or exam-style Q/A formatting.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
     पालिका
    0.42
     posso
    0.40
     fels
    0.38
    bfc
    0.38
     remem
    0.38
    াহিয়ার
    0.37
     onların
    0.37
     formato
    0.37
    िसोदिया
    0.37
    ＾＾
    0.37
    POSITIVE LOGITS
    Question
    0.57
     Question
    0.56
    What
    0.53
     question
    0.47
    Given
    0.46
     `
    0.45
    QUESTION
    0.44
     pregunta
    0.43
    Which
    0.43
    `
    0.43
    Act Density 0.087%

    No Known Activations