3 ms·Training language models to follow instructions with human feedback [pdf]3 points by josefslerka 4y ago