subnormal number
A float too small to be normalized, stored without the hidden 1 so values can shrink gradually toward zero.
A subnormal number has an all-zeros exponent field and a nonzero fraction, with value (-1)^s * 0.f * 2^(1 - bias); subnormals fill the gap between zero and the smallest normal number with evenly spaced values.
Examples
The smallest positive binary32 subnormal is 2⁻¹⁴⁹, about 1.4 × 10⁻⁴⁵: the pattern
0x00000001.