1 paper · 1 filter
Junyan Zhang, Yubo Gao, Yibo Yan +8
The finetuning of Large Language Models (LLMs) has significantly advanced their instruction-following capabilities, yet the underlying computational mechanisms driving these improv…