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

    Neuronpedia

    INDEX
    Explanations

    Chemical compounds

    np_max-act · gemini-2.0-flash

    chemical compounds and their safety-related properties.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8

    The neuron spikes on tokens that are part of complex chemical compound names (IUPAC-style nomenclature with numbers, hyphens, and element prefixes).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    Recipe
    -0.06
    检
    -0.06
     Elephant
    -0.06
    .MiddleRight
    -0.06
                                                ↵
    -0.06
    CCC
    -0.06
     hopeless
    -0.06
    Signed
    -0.06
    ается
    -0.06
    	container
    -0.06
    POSITIVE LOGITS
    _NONNULL
    0.08
    isFunction
    0.07
    аци
    0.07
    ξεις
    0.07
    ůj
    0.07
     mấy
    0.06
     kal
    0.06
    τρο
    0.06
    �
    0.06
     navegador
    0.06
    Act Density 0.048%

    No Known Activations