IEEE 754 float32 values in [0, 1]

Every glowing dot is a real single-precision float created the same way the hardware does: (1 + mantissa) × 2exponent.

Linear scale – the true number line. Almost every possible value sits extremely close to zero (the bright bar).
Log scale – spreads the values by order of magnitude so you can see that floats exist at every scale from 10⁻³⁸ up to 1.

WebGPU ready · 0 points
0
1
10⁻³⁰
10⁻²⁰
10⁻¹⁰
10⁻⁵
0.01
0.1