Cracking the Code of AI Alignment: This AI Paper from the University of Washington and Meta FAIR Unveils Better Alignment with Instruction Back-and-Forth Translation