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

    Neuronpedia

    INDEX
    Explanations

    article

    np_max-act · gemini-2.0-flash

    The neuron is triggered by positive descriptive adjectives and qualifiers (e.g. “just,” “pretty,” “good”) that add praise or approval.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    励
    -0.07
    alah
    -0.07
    .lr
    -0.06
    -brand
    -0.06
    	org
    -0.06
    .act
    -0.06
    "},
    -0.06
     Text
    -0.06
    _lat
    -0.06
     SWT
    -0.06
    POSITIVE LOGITS
    _PAGES
    0.07
    ÅŸ
    0.07
     esl
    0.06
     perman
    0.06
    어나
    0.06
     ऐ
    0.06
    억
    0.06
    Inset
    0.06
    ール
    0.06
     křes
    0.06
    Act Density 0.001%

    No Known Activations