7 papers
Dense Process Supervision for Search Agents via Fact Utility Estimation
Rongzhi Zhu, Xiangyu Liu, Yi Liu +7
Reinforcement learning (RL) for search agents typically relies on outcome rewards. However, it often fails to achieve effective credit assignment, due to the unclear value of inter…
Polished but Unresolved: Identifying Late-Stage Pressure States in Long-Horizon Tool-Use Agents
Haoyang Chen, Yi Liu, Jianzhi Shao +3
Long-horizon tool-use agents need not only to search and plan, but also to decide when to finalize. We study late-stage pressure states, in which an agent is biased toward submitti…
ActMem: Bridging the Gap Between Memory Retrieval and Reasoning in LLM Agents
Xiaohui Zhang, Zequn Sun, Chengyuan Yang +3
Memory management is essential for LLM agents in long-term interactions. Current memory frameworks typically treat agents as passive ``recorders'' and retrieve information without…
Answering the Unanswerable Is to Err Knowingly: Analyzing and Mitigating Abstention Failures in Large Reasoning Models
Yi Liu, Xiangyu Liu, Zequn Sun +1
Large reasoning models (LRMs) have shown remarkable progress on complex reasoning tasks. However, some questions posed to LRMs are inherently unanswerable, such as math problems la…
When Can Large Reasoning Models Save Thinking? Mechanistic Analysis of Behavioral Divergence in Reasoning
Rongzhi Zhu, Yi Liu, Jiancheng Wang +6
Large reasoning models (LRMs) have achieved remarkable success on complex tasks, yet their tendency to "overthink" leads to inefficiencies. Although "save-thinking" prompts are int…
Mitigating Lost-in-Retrieval Problems in Retrieval Augmented Multi-Hop Question Answering
Rongzhi Zhu, Xiangyu Liu, Zequn Sun +2
In this paper, we identify a critical problem, "lost-in-retrieval", in retrieval-augmented multi-hop question answering (QA): the key entities are missed in LLMs' sub-question deco…