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

    Neuronpedia

    INDEX
    Explanations

    physics and math equations

    np_max-act · gemini-2.0-flash

    The neuron activates on tokens that are part of in‐line mathematical content—numbers, equations, and TeX math delimiters.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    job
    -0.07
    부터
    -0.07
     entre
    -0.07
     planning
    -0.07
     Frank
    -0.07
    Besides
    -0.07
    Stuff
    -0.06
    South
    -0.06
     deputies
    -0.06
    ╗
    -0.06
    POSITIVE LOGITS
     Вік
    0.07
    \x
    0.07
    \S
    0.07
    \s
    0.07
    \d
    0.07
     ><?
    0.06
    (itemView
    0.06
     �
    0.06
     захист
    0.06
    .formData
    0.06
    Act Density 0.015%

    No Known Activations