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

    Neuronpedia

    INDEX
    Explanations

    Explanations and contextual notes

    np_max-act · gemini-2.0-flash

    The neuron activates on uppercase abbreviations or acronyms (e.g., EHR, GIL) in the text.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Capability
    -0.06
    uminum
    -0.06
    -inflammatory
    -0.06
     eerie
    -0.06
    ----------------------------------------------------------------------↵
    -0.06
    辅
    -0.06
     irgend
    -0.06
    (tex
    -0.06
     Optionally
    -0.06
     comeback
    -0.06
    POSITIVE LOGITS
     charms
    0.08
    erving
    0.07
    ,char
    0.07
     recognizing
    0.07
    atti
    0.07
    heid
    0.06
    PosX
    0.06
    (define
    0.06
    crate
    0.06
    처럼
    0.06
    Act Density 0.055%

    No Known Activations