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

    Neuronpedia

    INDEX
    Explanations

    be

    np_max-act · gemini-2.0-flash

    The neuron flags interrogative sentences (i.e. questions), activating on question‐mark punctuation and the auxiliary/verb patterns that form a question.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     compute
    -0.08
    bies
    -0.07
     변경
    -0.07
     train
    -0.07
     rooms
    -0.06
     ศร
    -0.06
     Champ
    -0.06
     ambiguity
    -0.06
    ("")↵
    -0.06
     get
    -0.06
    POSITIVE LOGITS
     στι
    0.07
     CLOSED
    0.06
    튜
    0.06
     tan
    0.06
    야
    0.06
     dj
    0.06
    STITUTE
    0.06
    setMax
    0.06
     YouTube
    0.06
    _ERRORS
    0.06
    Act Density 0.058%

    No Known Activations