INDEX
    Explanations

    research paper sections

    This neuron activates on LaTeX section markers and references (e.g. “sec:…,” “Section […]”) indicating it detects section headings and labels.

    New Auto-Interp
    Negative Logits
    Yesterday
    -0.06
    ्न
    -0.06
    seite
    -0.06
    shan
    -0.06
     mast
    -0.06
    udent
    -0.06
    PerPixel
    -0.06
    -0.05
     preval
    -0.05
    nění
    -0.05
    POSITIVE LOGITS
    .dist
    0.07
     altre
    0.07
    .accept
    0.07
     Website
    0.06
    (jq
    0.06
    (call
    0.06
     Elemental
    0.06
    elite
    0.06
     Teresa
    0.06
     numa
    0.06
    Act Density 0.003%

    No Known Activations