chore: replace 64bit keyword with 32bit in @stdlib/constants/float32/num-significand-bits

Package returns the number of significand bits in single-precision
(23) and is in the float32 namespace, but its keyword array contained
`64bit` alongside `float32`. Replace with `32bit` so the keywords
match the package's actual precision.

Author: claude

lib/node_modules/@stdlib/constants/float32/num-significand-bits/package.json

@@ -57,7 +57,7 @@
     "math",
     "float",
     "float32",
-    "64bit",
+    "32bit",
     "floating-point",
     "ieee754",
     "significand",