collaborators

7 papers

cs.AI2026

Retry, Switch, or Abstain? Learning Strategy-Aware Tool-Use Policies via Controlled Error Injection

Chaoran Chen, Vy Nguyen, Ziji Zhang +7

Tool-using LLM agents are commonly trained and evaluated in environments where tool calls succeed reliably, yet deployed tools can fail transiently, persistently, or silently. Robu…

cs.LG2026

Adaptive Layerwise Perturbation: Unifying Off-Policy Corrections for LLM RL

Chenlu Ye, Xuanchang Zhang, Yifan Hao +6

Off-policy problems such as policy staleness and training--inference mismatch have become a major bottleneck for training stability and further exploration in LLM RL. The distribut…

cs.LG2026

Beyond Correctness: Harmonizing Process and Outcome Rewards through RL Training

Chenlu Ye, Zhou Yu, Ziji Zhang +5

Reinforcement Learning with Verifiable Rewards (RLVR) improves final-answer accuracy on reasoning tasks, but it does not reliably improve reasoning quality. Because outcome rewards…

cs.CL2025

TOD-ProcBench: Benchmarking Complex Instruction-Following in Task-Oriented Dialogues

Sarik Ghazarian, Abhinav Gullapalli, Swair Shah +4

In real-world task-oriented dialogue (TOD) settings, agents are required to strictly adhere to complex instructions while conducting multi-turn conversations with customers. These…

cs.CL2025

Semantic Volume: Quantifying and Detecting both External and Internal Uncertainty in LLMs

Xiaomin Li, Zhou Yu, Ziji Zhang +4

Large language models (LLMs) have demonstrated remarkable performance across diverse tasks by encoding vast amounts of factual knowledge. However, they are still prone to hallucina…

cs.CL2025

When Thinking Fails: The Pitfalls of Reasoning for Instruction-Following in LLMs

Xiaomin Li, Zhou Yu, Zhiwei Zhang +5

Reasoning-enhanced large language models (RLLMs), whether explicitly trained for reasoning or prompted via chain-of-thought (CoT), have achieved state-of-the-art performance on man…