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

    Neuronpedia

    INDEX
    Explanations

    instructions and technical writing

    np_max-act · gemini-2.0-flash

    logical inconsistencies in claims about the presence of wildfire smoke or flames in images.

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

    This neuron activates on text where the model refers to its own reasoning—especially phrases like “your thought process.”

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     enf
    -0.08
     menuItem
    -0.07
    Hold
    -0.06
     signUp
    -0.06
     vazgeç
    -0.06
     promotes
    -0.06
     meat
    -0.06
     Meat
    -0.06
    ],
↵
    -0.06
    Ingredients
    -0.06
    POSITIVE LOGITS
    (currency
    0.07
    operations
    0.07
    且
    0.07
     poco
    0.07
    inars
    0.07
     difficulty
    0.06
    locator
    0.06
     Voyage
    0.06
    гляд
    0.06
    /components
    0.06
    Act Density 0.006%

    No Known Activations