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

    Neuronpedia

    INDEX
    Explanations

    go

    np_max-act · gemini-2.0-flash

    This neuron activates on the instruction prompt requiring the answer to begin explicitly with “Yes” or “No.” It detects the directive about how to format the response (starting with "Yes" or "No").

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Tah
    -0.06
    viso
    -0.06
    \Test
    -0.06
     ++;↵
    -0.06
     این
    -0.06
    enou
    -0.06
     printk
    -0.06
     Boehner
    -0.05
    (and
    -0.05
    자의
    -0.05
    POSITIVE LOGITS
    uring
    0.07
     converged
    0.07
     مورد
    0.07
     unidentified
    0.07
     lending
    0.07
    -local
    0.07
    slide
    0.07
     squirrel
    0.07
    RELATED
    0.07
    -party
    0.06
    Act Density 0.008%

    No Known Activations