My guess is that the mutations are retained as variation in the population because having some small subset of them probably results in divergent thinking which can increase overall group survival. Autistic spectrum disorder individuals result from having too many of these mutations, such that neural activity patterns find a stable equilibrium outside of "normal" (think strange attractors in system dynamics).

A reasonable theory, but as a minor note most evolutionary biologists wouldn't focus on group survival. Group selection remains controversial.

But your point still stands, and has clear parallels in other cases, like sickle-cell anemia: one copy of the gene is helpful, so it remains in the population, but two copies is harmful, so its spread is limited.

I'm not an evolutionary biologist, but this article (https://en.wikipedia.org/wiki/Evolutionary_game_theory) mentions Hamilton's rule: A given altristic gene will propogate if "the genetic relatedness of the altruistic donor exceeds the cost-benefit ratio of the altruistic act itself". Is this not widely agreed upon?

Evolution isn't directed.

It sounds as though you are speaking from the standpoint that genes mutate at random in all and any "directions" (net). However, non-successful genes are eliminated, so in the effective sense it is (gross). The direction of evolution is towards being "more suitable" for a particular environment.

And if there's no pressure to remove a particular trait from being in the reproductive environment...

Either genes survive or they don't. There's no forwards or backwards.

backward: these traits are a disadvantage to host organism. forward: they are an advantage, and thus survive.

Are they aligned with natural selection processes in the environment.

evolution doesn't care about organisms or species, only genes.

I'm not sure why you're anthropomorphising evolution. It's a process, not even a thing. It's a consequence of interacting environmental conditions, and it doesn't "care" at all in any sense of the word. No more than point-charges "care" about electromagnetism.

dragonwriter is right in general sense. We can evaluate it, though, from our own perspective. It's a step forward in terms of intelligence and stability in world. It's a giant leap backward in terms of success or happiness in a world of non-autistic people. Evolution would also consider it a negative in that autistics get laid and reproduce less than non-autistics. They also are more likely to loose in competition in corporations, battlefields, and jungles. So, it's a net loss overall for evolution's main function but autistics persist for one reason or another.

There's a lot of rampant speculation in the comment here.

Just opinions like most interpretations in that field of study. That's all.