One of the main problems for evolving intelligence is that it seems to be heavily learning-based, and antagonistic to instinct. For a variant with more of a learning style to succeed, it must go from being a high-instinct type where it knows everything automatically, to not automatically knowing things and having to learn them. "Not automatically knowing something" is fairly easy to evolve from "automatically knowing something", but having to go through some process just to catch up on knowledge its competitors had from birth is a tricky thing to do while still enjoying some competitive advantage. I think that's why high intelligence evolves rarely and slowly.
Benefitting from learning works best with learning by copying since in the process of trial and error the trials are more likely to be useful, so parental protection and long-term attachment would seem to be advantageous. I agree with John Bois that vivipary is important along with large brain size and long "childhood".
Having a lifestyle suitable for learning would make learning tendencies more likely to succeed of course, and in man's recent history tools were of enormous significance; however tools don't explain the encephalisation of his earlier ancestors nor that of dolphins, so . . . to get back to dinos, tool-manipulating hands would not be necessary for intelligence.
I don't see why Troodons or Oviraptors couldn't have had an environment with enough potential flexibility to make a high learning style pay off. However, having a long neck and catching small active prey in your mouth might work against evolving a large brain (though mouth-catching didn't hold the dolphins up). An animal that could make its way both on the ground and in trees, and could eat vegatable and large and small animal material, catching it in its mouth, hands or feet, would seem to have the best chances of having the wide variety of possibilities that would work well with a high learning lifestyle.
Did you know, there is a big african parrot that goes to the top of a tree and displays by whacking the tree with a stout stick it holds in one foot. It even seems to have a possibly humanlike hairstyle. It seems to me it is mimicking a human; however since the parrot is black, this might mean it had to have evolved this in the last 50,000 years. Incidentally, parrots have particularly squidgey foot pads don't they? (not to mention zygo tendencies) which would pre-adapt them to wield tools.