INDEX
    Explanations

    technical writing

    The neuron activates on words that denote foundational concepts—specifically “basic” and “principles.”

    New Auto-Interp
    Negative Logits
    .ci
    -0.07
    _One
    -0.07
    Propagation
    -0.07
    uyệ
    -0.06
     pare
    -0.06
    alers
    -0.06
    ้ด
    -0.06
     overseas
    -0.06
    унок
    -0.06
    sizlik
    -0.06
    POSITIVE LOGITS
    ."
    0.07
    .”
    0.07
    )."
    0.06
     ”↵
    0.06
    ","
    0.06
    ,',
    0.06
    )/(
    0.06
    .”↵↵
    0.06
     "),↵
    0.06
    "}
    0.06
    Act Density 0.000%

    No Known Activations