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

    Neuronpedia

    INDEX
    Explanations

    evaluate

    np_max-act · gemini-2.0-flash

    This neuron detects discussion of model evaluation steps—especially splitting data into test sets and evaluating performance.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    Dto
    -0.07
    >Description
    -0.06
    iplinary
    -0.06
     bagi
    -0.06
    _INVALID
    -0.06
     либо
    -0.06
     Blasio
    -0.06
    	In
    -0.06
     ComponentFixture
    -0.06
    认为
    -0.06
    POSITIVE LOGITS
     nw
    0.08
     gint
    0.07
     прот
    0.07
     testing
    0.07
     bước
    0.07
    drops
    0.06
     Ř
    0.06
     سب
    0.06
    ipse
    0.06
     سک
    0.06
    Act Density 0.015%

    No Known Activations