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

    Neuronpedia

    INDEX
    Explanations

    punctuation marks

    np_max-act · gemini-2.0-flash

    The neuron flags tokens that are part of the assistant’s (AI’s) response segments, i.e. it detects when the text is coming from the assistant rather than the user.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    ades
    -0.06
    instances
    -0.06
    ạm
    -0.06
     heir
    -0.06
    questions
    -0.06
    .enabled
    -0.06
     Rates
    -0.06
     düşman
    -0.06
    radio
    -0.06
    �
    -0.06
    POSITIVE LOGITS
     及
    0.07
     QVERIFY
    0.07
    КТ
    0.07
    roveň
    0.07
    :@{
    0.07
    )/(
    0.06
    及
    0.06
     bied
    0.06
    ाहत
    0.06
     smooth
    0.06
    Act Density 0.035%

    No Known Activations