No, a chimeric virus created in a gain-of-function experiment would look extremely similar to known viruses, because these chimeras are created by combining elements of known viruses. It would not be 4% different from the closest known natural virus. Accumulating thousands of mutations throughout the entire genome takes decades of evolution. In the wild, that means thousands of generations of hosts.

A virus created in a gain-of-function experiment would also use a well-known backbone. It would not be based on some virus that nobody had ever heard of.

The problem with that theory is that the virus contains thousands of mutations throughout its genome, which would take decades of evolution to accumulate. In the wild, this virus jumps to a new host every few days, meaning that decades of evolution amount to thousands of generations of virus. No lab has the time to pass virus through so many animals.