If I was a potential employer looking at the results, I would disqualify a job candidate that cheated like this.

My solution passed all the tests but failed with a backtick in the only the second position. I've fixed it locally, and updated my solution, but only the first solution is showing up for me.

It is the exception where you can encode 14 bits instead of 13 (For values 8192 - 8280). This throws off the bits of the following characters and will give you the results you found. If you put an '@' or '`' in the second position it will happen.

The instructions are somewhat confusing.

I had trouble finding reference material, and only by reading pre-existing source code did I figure out that I was reading the characters in the wrong direction, and you need to pull the 13/14 bits from the least significant bits, not quite how I understand Base64 works.

For example: 'H','e' in binary is 01001000 01100101... The first 13 bits would be 0100100001100 which encodes to "pZ". The algorithm to produce a result that starts with ">O" needs the byte order to be reversed to 01100101 01001000 and taking the 13 rightmost bits: 0010101001000 or 78d 14d.

Biggest issue is that the random tests are not random, just testing against a sequence of "%s".

Sad to see that most of the solutions are copies of Joachim Henke's C# source code from Codeplex. Some just changed the variable names he used (b, n, v) to more meaningful ones (queue, bits, val); some just changed from string to stringbuilder; some just added a completely superfluous "and 255" operation to an array of bytes...

This is more an exercise of research skills than coding skills.