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

    Neuronpedia

    INDEX
    Explanations

    Mathematical equations

    np_max-act · gemini-2.0-flash

    The neuron activates on parts of the text that are inside or immediately adjacent to LaTeX‐style mathematical expressions (e.g. math mode markers, variables like x, y, operators like ∂, frac, etc.).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     che
    -0.06
    tones
    -0.06
     Tak
    -0.06
     categorized
    -0.06
     enable
    -0.06
     Rahman
    -0.06
    วด
    -0.06
    439
    -0.06
    :])
    -0.05
     tratt
    -0.05
    POSITIVE LOGITS
    .inf
    0.08
    layın
    0.07
    (stdin
    0.07
    inion
    0.07
     Ả
    0.07
    107
    0.07
    .dst
    0.06
     spokesperson
    0.06
     gdy
    0.06
     TimeUnit
    0.06
    Act Density 0.033%

    No Known Activations