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

    Neuronpedia

    INDEX
    Explanations

    }

    np_max-act · gemini-2.0-flash

    This neuron detects structural section headings and formatting markers (e.g. “Experimental,” “Crystal data,” numbered subsections, “###,” “====,” etc.) in scientific‐paper–style text.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Rocky
    -0.07
     Hatch
    -0.07
    -divider
    -0.07
    ��
    -0.07
     webs
    -0.07
    .called
    -0.06
     стра
    -0.06
    Inlining
    -0.06
    ]>
    -0.06
    Latch
    -0.06
    POSITIVE LOGITS
     learner
    0.06
    _caption
    0.06
     PRODUCTS
    0.06
    udi
    0.06
    _media
    0.06
    Pear
    0.06
    .tv
    0.06
    XP
    0.06
     dopo
    0.06
    .TEXT
    0.06
    Act Density 0.001%

    No Known Activations