Xiaoqing Wu

Misleading historical data can corrupt over 30% of tool-calling decisions, but a new method can restore accuracy by effectively transferring reliable policies from teacher to student models.

Evidence recall, not precision, is the unsung hero of QA systems, explaining 96% of answer quality variance.

LLM-Wiki shows that structuring external knowledge as a self-evolving Wiki, rather than a flat index, unlocks significant gains in LLM reasoning and multi-document query accuracy.