← Single precision floating-point format