What is the difference between ILP and thread-level parallelism?

My understanding is that superscalar parallelism (I guess what you meant by ILP here) is when we have one core but multiple ALUs, we figure out the independent instructions (they are from the same thread) to run simultaneously on these different ALUs. Thread-level parallelism is when we have multiple cores, each of them can run an entirely different stream of instructions (or thread).