3 papers
cs.CL2026
ToolLoop: Closed-Loop Tool-Use Data Synthesis via Decomposed Generation and Dynamic Self-Feedback
Min Zeng, Yuzhou Liu, Zhenyu Cao +5
High-quality tool-use data is critical for training language models to interact effectively with external tools. However, existing synthetic approaches typically follow a generate-…
cs.CL2026
Learning Simple Test-Time Environments for LLM Web Agents
Junxuan Li, Zijun Liu, Ziyi Huang +4
Large language model (LLM) agents have demonstrated remarkable proficiency in manually constructed environments, yet their performance frequently collapses when transitioned to com…
cs.CL2026
Cloud-ScPO: Hidden-State Geometry for Semi-Supervised Preference Optimization in LLM Reasoning
Yuzhou Liu, Xiyang Hu
Preference optimization improves mathematical reasoning in large language models (LLMs), but reliable chosen-rejected pairs usually require verified answers, human annotations, or…