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

    Neuronpedia

    INDEX
    Explanations

    Pat

    np_max-act · gemini-2.0-flash

    references to patents or patent-related concepts.

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

    The neuron responds to tokens beginning with “Pat,” i.e. the substring “Pat” at the start of a word.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Greenwood
    -0.07
    雲
    -0.07
    2
    -0.06
    为空
    -0.06
     Wein
    -0.06
     Wine
    -0.06
     easy
    -0.06
     opened
    -0.06
    66
    -0.06
    Simple
    -0.06
    POSITIVE LOGITS
     Pat
    0.14
     Patricia
    0.13
    Pat
    0.12
     pat
    0.10
     patrol
    0.10
    (pat
    0.10
     Patrick
    0.09
     Patrol
    0.09
     PAT
    0.08
    pat
    0.08
    Act Density 0.012%

    No Known Activations