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

    Neuronpedia

    INDEX
    Explanations

    code

    np_max-act · gemini-2.0-flash

    The neuron strongly activates on tokens that specify measurement units (e.g. “unit,” “cup,” “teaspoon,” “pkg,” “count,” etc.) in an ingredients list.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    phrases related to recipes or cooking instructions.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    Opening
    -0.07
     herbal
    -0.06
    �
    -0.06
    -0.06
    _datas
    -0.06
     glass
    -0.06
     قرارداد
    -0.06
     дер
    -0.06
    智能
    -0.06
     commun
    -0.06
    POSITIVE LOGITS
     textile
    0.07
     будет
    0.06
    Indices
    0.06
    _sv
    0.06
    	sleep
    0.06
     første
    0.06
    ілля
    0.06
    _quota
    0.06
    step
    0.06
    .AppendFormat
    0.06
    Act Density 0.002%

    No Known Activations