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

    Neuronpedia

    INDEX
    Explanations

    large language model

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

    references to the model's identity or the phrase "As a large language model" (self‑referential model introductions).

    oai_token-act-pair · gpt-5-miniTriggered by @yooniel31

    The neuron activates on the self‐referential “As a large language model” style disclaimer phrase.

    oai_token-act-pair · o4-miniTriggered by @yooniel31
    New Auto-Interp
    Negative Logits
     convirt
    0.42
     sehingga
    0.40
     wodurch
    0.40
     بنابراین
    0.37
    。
    0.37
    に示す
    0.37
     випад
    0.37
     تركيب
    0.36
     ngunit
    0.36
     ପ୍ର
    0.35
    POSITIVE LOGITS
     indexRouter
    0.43
     étant
    0.42
    我都
    0.40
     being
    0.39
     YouTuber
    0.39
    having
    0.39
     having
    0.38
     itself
    0.38
    我会
    0.38
     lover
    0.37
    Act Density 0.069%

    No Known Activations