-Crichton's "raptors" were actually supposed to be the giant Asian dromaeosaurid that would eventually be named Achillobator, which was mentioned in Predatory Dinosaurs Of The World, the book that inspired JP's dinosaur descriptions. The remains weren't officially described until 1999, but Gregory Paul apparently had some conception of the size 11 years beforehand.
The film used Deinonychus as the visual basis, as it was the largest well-known dromaeosaur at the time, then extrapolated a more heavyset creature from that.
As for the name, Crichton subscribed to a theory in the '80s that all dromaeosaurids were so closely related that they should be united under one genus - Velociraptor, being the oldest name.
-JP's rex is slightly bigger than the largest known Tyrannosaurus, Sue, but it's within theoretical bounds for size variance within a species. Given that it's only known from 50 individuals, it's a pretty safe bet that there were adults far smaller and larger than the ones we've found.
As for the 'visual accuity based on movement' - believe it or not, that's based in science too. Crichton liked to showcase cutting edge research, and an emerging field during the '80s was the analysis of fossilised braincases to draw parallels to living animals. Early research suggested Tyrannosaurus' sensory centre was similar to that of an amphibian, and Crichton extrapolated from that.
Of course, we now know that the exact opposite was true, and Tyrannosaurus actually had exceptional eyesight, but the science has come a lot further since then.
-Dilophosaurus is of course a big bundle of wrong, in the film at least. As mentioned before, the venom and frill were speculative. As for the size and head shape, this was an honest mistake, based on the designer using the original holotype for reference -
(For anyone who doesn't know, Dilophosaurus was originally thought to be a new species of Megalosaurus.)
This was actually a very common mistake, and my pre-JP dinosaur books also showed Dilophosaurus with a generic carnosaur skull, rather than the pointy, long-toothed coelophysoid skull it actually had.
-All the theropods have pronating 'bunny hands'. We now know that prosauropods and theropods had very inflexible wrists and forearms, particularly with regard to the position of the hands - in other words, holding a basketball, not dribbling it. The dinosaurs right at the avian root (such as Velociraptor) had additional bones which allowed them to bend the hand sideways towards the wrist, but even they couldn't make their palms face the floor without splaying their arms close to 90 degrees.
However, this is based on studies that were published between 2002 and 2008, so even Jurassic Park 3 wasn't inaccurate for its time, in this regard.
-Feathers. We now know that all coelurosaurs - the group that includes Compsognathus, Gallimimus, Tyrannosaurus and Velociraptor - had proto-feathers as the primary integument. It's important to note that 'proto-feather' is a broad-ranging term - compsognathid feathers were very similar to mammalian hair, ornithomimosaurs had fluffy down like a bird chick, dromaeosaurids had proper wing feathers.
Tyrannosaurus itself is probably safe though - multi-ton animals from warm/temperate climates don't need insulation, so if it had any kind of feather, it would have been for display. My guess is that it was scaly all over, and we have found samples of scaly skin in its closest relatives. That said, rex junior would probably have been massively fluffy and cute as a button.
Anyway. The evidence for feathers in dinosaurs didn't really reach a tipping point until Microraptor was described in 2002, after even Jurassic Park 3.
-Finally, cloning dinosaurs from mosquitos trapped in amber. When Crichton wrote Jurassic Park, it was genuinely believed in the scientific community that this was possible. Partial strands of DNA had recently been found in Cretaceous amber, and it was thought it was just a case of finding the right amber, and achieving the understanding of genetics to be able to 'patch' missing base pairs (the vast majority of any genome consists of 'junk' pairs anyway).
Crichton addresses the various other difficulties of cloning a dinosaur in the book.
Anyway, it later turned out that at least
some of that Cretaceous DNA was actually a more recent contaminant (there's still a piece of plant matter that my cursory search didn't find any evidence disproving). We've only ever found a couple of mosquitos preserved in Mesozoic amber, and even if we could find one that had recently drunk dinosaur blood, it seems there would almost inevitably be contaminants from the bug itself, even if any of the DNA was preserved, and even if we could figure out how to patch the genome.
Still. This has largely been figured out due to the scrutiny that was placed on the subject after Crichton's writing.