I left some comments on #275 - I would prefer to add an additional lerp_precise method than modify the existing lerp method. Primarily because the existing lerp method can be used to extrapolate and if people are using it that way this would break their code. A separate method means the precise option is there for those who need or want it.

Documentation would need to be added for the new and old methods explaining the difference.

It would also be good to add glam_assert checks to ensure t is in a valid range for the precise version.