10 papers
When History Lies: Evaluating and Improving Tool Use under Misleading Multi-Turn Histories
Xiaoqing Wu, Xingyu Fan, Feifei Li +1
Tool-calling agents infer task state from accumulated dialogue and tool traces. In persistent interactions, however, historical traces may remain structurally valid and semanticall…
DUET: Dual-Teacher On-Policy Distillation via Same-Weight Disagreement for Prohibition Compliance
Zihan Li, Feifei Li, Wenhui Que
Real-world LLM deployments increasingly rely on runtime-injected prohibitions--enterprise policies, PII redlines, tool boundaries--that vary per request and per tenant. Conventiona…
WikiLoop: Jointly Learning to Build and Navigate Agent-Native Wikis with Downstream Feedback
Haoliang Ming, Feifei Li, Wenhui Que
WikiLoop is a framework that jointly learns to build a machine‑readable wiki and to navigate it for answering queries, using downstream answer correctness as feedback to guide stru…
WikiKV: Schema-Evolving Path-Indexed Storage for Hierarchical Knowledge Navigation
Feifei Li, Haoliang Ming, Zihan Li +5
LLM-curated hierarchical knowledge bases, namely a tree-structured wiki whose nodes summarize an underlying corpus, have become a dominant substrate for retrieval-augmented applica…
MemCog: From Memory-as-Tool to Memory-as-Cognition in Conversational Agents
Zihan Li, Xingyu Fan, Feifei Li +1
Existing agent memory systems universally follow what we term a Memory-as-Tool paradigm where a single query triggers one-shot retrieval of flat passage lists, suffering from passi…
Semantic Flow Regularization: Teaching LLMs to Generate Diverse Yet Coherent Responses
Kerui Peng, Feifei Li, Xingyu Fan +1
When large language models are fine-tuned to generate persona- or tone-conditioned responses, their output diversity is severely limited--a failure we term Cross-Style Collapse. We…