Tech--多轮对话 RL

• RAGEN / StarPO: https://arxiv.org/abs/2504.20073

• MUA-RL: https://arxiv.org/abs/2508.18669

• UGST / Goal Alignment: https://arxiv.org/abs/2507.20152

• SimulatorArena: https://arxiv.org/abs/2510.05444

• SAGE: https://arxiv.org/abs/2510.11997

• Persona Simulator RL: https://arxiv.org/abs/2511.00222

• SWEET-RL / ColBench: https://arxiv.org/abs/2503.15478

• Turn-Level Rewards: https://arxiv.org/abs/2505.11821

• τ2-bench: https://arxiv.org/abs/2506.07982

• AgentGym-RL: https://arxiv.org/abs/2509.08755

• AgentRL: https://arxiv.org/abs/2510.04206

• Agent-R1: https://arxiv.org/abs/2511.14460

• RAGEN-2: https://arxiv.org/abs/2604.06268

• TSR: https://arxiv.org/abs/2602.11767

• ProRL Agent: https://arxiv.org/abs/2603.18815

• LOOP / Long-Horizon Interactive LLM Agents: https://arxiv.org/abs/2502.01600

• RAGEN GitHub: https://github.com/RAGEN-AI/RAGEN

• AgentRL GitHub: https://github.com/THUDM/AgentRL

• Agent-R1 GitHub: https://github.com/AgentR1/Agent-R1

• AgentGym-RL project: https://AgentGym-RL.github.io

• veRL: https://github.com/volcengine/verl

• ROLL: https://github.com/alibaba/ROLL

• SkyRL: https://sky.cs.berkeley.edu/project/skyrl/