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

    Neuronpedia

    INDEX
    Explanations

    code special characters

    np_max-act · gemini-2.0-flash

    This neuron detects command‐line option tokens (i.e. short or long flags beginning with “-” or “--” such as -v, -h, --help, --dry-run).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    #[
    -0.07
    inx
    -0.07
    Routing
    -0.06
     Cres
    -0.06
    multiline
    -0.06
    					↵					↵
    -0.06
    _parms
    -0.06
    ificados
    -0.06
     Casc
    -0.06
    .↵↵↵↵↵
    -0.06
    POSITIVE LOGITS
    ční
    0.07
     Permanent
    0.06
     tracing
    0.06
    cle
    0.06
     zboží
    0.06
     bonne
    0.06
    íd
    0.06
    .va
    0.06
    blo
    0.06
     @"\
    0.06
    Act Density 0.006%

    No Known Activations