8 papers
Confidence-Calibrated Inference Expansion for Evaluator-Guided Test-Time Reasoning
Weida Liang, Kenji Kawaguchi
Evaluator-guided reasoning presents a local control problem at every step: trust the current answer, return a fallback, or spend more inference. We formulate this choice as a compu…
Bayesian Test-Time Inference of Task-Aligned Similarity from Weak Interactive Feedback
Weida Liang, Kenji Kawaguchi
Language model systems repeatedly make test-time decisions by comparing similarity: they retrieve exemplars, rank candidates, and revise responses. Yet similarity is not unique: le…
MemPilot: Orchestrating On-Demand Multimodal Memory Curation for LLM Agents
Haozhen Zhang, Haodong Yue, Quanyu Long +6
Memory has become integral to the LLM agent ecosystem, supporting information retention and reuse across interactions. However, most existing agent memory systems construct memory…
OpenSkill: Open-World Self-Evolution for LLM Agents
Zhiling Yan, Dingjie Song, Hanrong Zhang +8
Self-evolving agents requires adaptation after deployment, but existing approaches assume a usable learning loop, such as curated skills, successful trajectories, or verifier signa…
Strategy Executability in Mathematical Reasoning: Leveraging Human-Model Differences for Effective Guidance
Weida Liang, Yiyou Sun, Shuyuan Nan +3
Example-based guidance is widely used to improve mathematical reasoning at inference time, yet its effectiveness is highly unstable across problems and models-even when the guidanc…
From Harm to Help: Turning Reasoning In-Context Demos into Assets for Reasoning LMs
Haonan Wang, Weida Liang, Zihang Fu +8
Recent reasoning LLMs (RLMs), especially those trained with verifier-based reinforcement learning, often perform worse with few-shot CoT than with direct answering. We revisit this…
Improving Conversational Recommendation with Contextual Adaptation of External Recommenders and LLM-based Reranking
Chuang Li, Weida Liang, Hengchang Hu +4
We tackle the challenge of integrating large language models (LLMs) with external recommender systems to enhance domain expertise in conversational recommendation (CRS). Current LL…
PromptArmor: Simple yet Effective Prompt Injection Defenses
Tianneng Shi, Kaijie Zhu, Zhun Wang +13
Despite their potential, recent research has demonstrated that LLM agents are vulnerable to prompt injection attacks, where malicious prompts are injected into the agent's input, c…