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

    Neuronpedia

    INDEX
    Explanations

    Ultra

    np_max-act · gemini-2.0-flash

    The neuron is primarily activated by the token “Ultra” (i.e. the “ultra” prefix).

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

    terms related to scientific measurements and structural studies.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Ezek
    -0.07
     homosex
    -0.07
     Fey
    -0.07
    =y
    -0.07
     Debbie
    -0.07
    emey
    -0.07
    _sent
    -0.07
    olkien
    -0.07
     Freund
    -0.07
     Manhattan
    -0.06
    POSITIVE LOGITS
     ultra
    0.12
     Ultra
    0.11
    Ultra
    0.10
     ultr
    0.10
     Ultr
    0.10
    ltr
    0.09
     XT
    0.08
    la
    0.08
    .Ultra
    0.08
    last
    0.07
    Act Density 0.011%

    No Known Activations