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

    Neuronpedia

    INDEX
    Explanations

    S

    np_max-act · gemini-2.0-flash

    This neuron detects isolated single-letter identifiers (e.g. “M”, “R”, “S”) typically used as enum variant or configuration option names in code.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    	diff
    -0.07
    rı
    -0.06
    TREE
    -0.06
    cccc
    -0.06
    prt
    -0.06
    lak
    -0.06
    brick
    -0.06
    stat
    -0.06
    ,min
    -0.06
    it
    -0.06
    POSITIVE LOGITS
    -S
    0.13
    -M
    0.11
    -H
    0.11
    .S
    0.11
    .H
    0.11
    .M
    0.11
    .R
    0.11
    -R
    0.10
    .T
    0.10
    .D
    0.10
    Act Density 0.397%

    No Known Activations