I noticed some strange behavior today with how Notepad++ handles encoding detection: if a file contains 1 or more section symbols (§) and no other non-ASCII characters, Notepad++ will decode the file using the TIS-620 charset, resulting in each section symbol being decoded to ยง. This behavior is entirely inconsistent and does not seem to occur for any other non-ASCII Latin character, and will not occur if other non-ASCII symbols are present in the file.

Edit: I should mention this occurs on multiple machines. all with an English locale active.