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

    Neuronpedia

    INDEX
    Explanations

    more

    np_max-act · gemini-2.0-flash

    The neuron responds to subordinate‐linking function words (like “once,” “since,” “after,” etc.) that introduce dependent clauses.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    against
    -0.07
     flames
    -0.07
    近
    -0.07
    اح
    -0.06
    それは
    -0.06
     شه
    -0.06
    _evaluation
    -0.06
    plans
    -0.06
     wei
    -0.06
    esus
    -0.06
    POSITIVE LOGITS
     노래
    0.07
     nemá
    0.06
    _IF
    0.06
     LOW
    0.06
    _DONE
    0.06
     Anda
    0.06
    _UINT
    0.06
     Alibaba
    0.06
     Nikon
    0.06
     indexOf
    0.06
    Act Density 0.137%

    No Known Activations