5 papers
CocoaBench: Evaluating Unified Digital Agents in the Wild
CocoaBench Team, Shibo Hao, Zhining Zhang +29
LLM agents now perform strongly in software engineering, deep research, GUI automation, and various other applications, while recent agent scaffolds and models are increasingly int…
RAGEN-2: Reasoning Collapse in Agentic RL
Zihan Wang, Chi Gui, Xing Jin +13
RL training of multi-turn LLM agents is inherently unstable, and reasoning quality directly determines task performance. Entropy is widely used to track reasoning stability. Howeve…
A Simple "Try Again" Can Elicit Multi-Turn LLM Reasoning
Licheng Liu, Zihan Wang, Linjie Li +5
Multi-turn problem solving is critical yet challenging for Large Reasoning Models (LRMs) to reflect on their reasoning and revise from feedback. Existing Reinforcement Learning (RL…
Online Prediction with Limited Selectivity
Licheng Liu, Mingda Qiao
Selective prediction [Dru13, QV19] models the scenario where a forecaster freely decides on the prediction window that their forecast spans. Many data statistics can be predicted t…
RAGEN: Understanding Self-Evolution in LLM Agents via Multi-Turn Reinforcement Learning
Zihan Wang, Kangrui Wang, Qineng Wang +15
Training large language models (LLMs) as interactive agents presents unique challenges including long-horizon decision making and interacting with stochastic environment feedback.…