L+M-24: Building a Dataset for Language+ Molecules