…Our code is available at https://github.com/hongruhou89/ProRL. View arXiv page View PDF GitHub 40 Add to collection Community Standard policy gradients are fundamentally broken for proactive recommendation. ProRL fixes…
…Through a closed-loop run--verify--reflect process , the framework jointly improves decomposition and execution over time via persistent, human-readable external memory , with self-evolving updates to each single-agent. During…
…On LoCoMo, EvolveMem outperforms the strongest baseline by 25.7% relative and achieves a 78.0% relative improvement over the minimal baseline. On MemBench, EvolveMem exceeds the strongest baseline by 18.9…
…Is there a source code available? Get this paper in your agent: hf papers read 2605.05806 Don't have the latest CLI? curl -LsSf https://hf.co/cli/install.sh | bash…
…Furthermore, improvements from model scaling quickly saturate, strong general reasoning does not reliably translate to creative affordance discovery, and common inference-time strategies such as Chain-of-Thought yield limited gains. These…
…Xiaoyu Yang , , , Abstract A novel framework called Autonomous Preference Optimization (APO) is proposed to address reasoning alignment challenges in multi-modal large language models under concept drift conditions, achieving improved robustness and…
…https://huggingface.co/zhuhz22/Causal-Forcing And the full-stack open-source code: https://github.com/thu-ml/Causal-Forcing We release 2-step frame-wise AR model with 50% latency and…
…Self-Evaluated Process Rewards for Retrieval-Augmented Agents (2026) Improving Search Agent with One Line of Code (2026) Enhancing LLM-based Search Agents via Contribution Weighted Group Relative Policy Optimization (2026) PRAISE…
…Generated by Qwen/Qwen2.5-Coder-32B-Instruct Larger models learn tasks smaller models do not. What drives this phenomenon? We develop a simple phenomenological argument that power-law scaling already suggests…
…Generated by Qwen/Qwen2.5-Coder-32B-Instruct We introduce CausaLab, a scalable environment for evaluating interactive causal discovery by LLM agents. Unlike prior evaluations, CausaLab evaluates both whether an agent can…