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

    Neuronpedia

    INDEX
    Explanations

    references/sources

    np_max-act · gemini-2.0-flash

    sections of text that indicate sources or references for information.

    oai_token-act-pair · gpt-4oTriggered by @xinyanhu8

    the main thing this neuron does is detect citation or reference language—phrases that attribute or source information (e.g. “References,” “Some of the information,” “Material from the Associated Press,” etc.).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    statements that attribute content to external sources or contributors, such as notices of copied/translated material, credits, and “contributed to this report”–style provenance.

    oai_token-act-pair · gpt-5Triggered by @liuliujiujiu05
    New Auto-Interp
    Negative Logits
    .gold
    -0.07
     chem
    -0.07
     switches
    -0.07
     فال
    -0.07
    612
    -0.07
    673
    -0.06
     rins
    -0.06
     Superior
    -0.06
    .dtp
    -0.06
    -yellow
    -0.06
    POSITIVE LOGITS
     conceivable
    0.06
    нист
    0.06
    =models
    0.06
    链接
    0.06
    _reward
    0.06
    ایه
    0.06
     KeyboardInterrupt
    0.06
     PropertyChanged
    0.06
    istické
    0.06
    ileceğini
    0.06
    Act Density 0.039%

    No Known Activations