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

    Neuronpedia

    INDEX
    Explanations

    comparisons, code

    np_max-act · gemini-2.0-flash

    financial metrics and terms related to profit and loss statements.

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

    The neuron detects comparative constructions—especially the “vs” token (and adjacent punctuation) that signals a contrast between two options.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    filt
    -0.07
    Decrypt
    -0.07
    璃
    -0.07
     Pad
    -0.06
     Wire
    -0.06
     HOUSE
    -0.06
    ANY
    -0.06
     pressures
    -0.06
     dan
    -0.06
    PRETTY
    -0.06
    POSITIVE LOGITS
    onomic
    0.06
     deriv
    0.06
    binary
    0.06
    熟
    0.06
    abcdefghijkl
    0.06
    boro
    0.06
     conformity
    0.06
    coop
    0.06
    ايد
    0.06
     cerv
    0.06
    Act Density 0.267%

    No Known Activations