Overflow occurs when the resulting value of an operation performed on valid representations of numbers is out of the range of valid values. That is, the resulting value is greater than max or less than min
If you are representing numbers using 4 bits as unsigned binary, the minimum value is 0 (i.e., 00002, while the maximum value is 15 (11112). Any operation that results in a number larger than 15 or smaller than 0 has overflowed.
For example, if you summed 14 + 15 (in UB, that's 1110 + 1111), the resulting value is 29, but the value 29 has no representation using 4 bits (since the maximum representable value is 15). Thus, there is overflow.
You might wonder, well, why not use 5 bits? Hardware has fixed number of bits. If it's 4 bits, then that's all you have to work with.
Detecting overflow in unsigned binary is straightforward. If adding the two values causes a N+1 bit result (this means that when you summed the MSb's plus any carry into the column of the MSBs, the resulting value had a carry to the next column), you have overflow.
https://www.cs.umd.edu/class/sum2003/cmsc311/Notes/Data/overflow.html