1 paper · 1 filter
Yichen Liu, Mingyu Chen, Hao Wang +7
Large language models are now adapted through chains of post-training stages rather than through a single instruction-tuning pass. This paper studies whether such sequential post-t…