An unfortunate reality of trying to represent continuous real numbers in a fixed space (e.g. with a limited number of bits) is that this comes with an inevitable loss of both precision and accuracy.

Once the domain of esoteric scientific and business computing, floating point calculations ... and LLMs do not care that much about float point precision, so the ~7.5% accuracy of the integer ...

New Bit and Cycle Accurate, Single, Double and Full Custom Precision Floating-Point Now Available in System Generator for DSP SAN JOSE, Calif., Nov. 1, 2011 -- Xilinx, Inc. (NASDAQ: XLNX) today ...