Tracked topic
…On the LunarLander-v2 delayed-reward benchmark, our decoupled agent avoids the "hovering for survival" local optimum. It eliminates policy collapse and stably surpasses the "Environment Solved" threshold without hyperparameter hacking. Code…
…every reasoning step — no parsing, no OCR, fully auditable. 📊 71.1% evidence-localisation accuracy on our new Wiki-CoE benchmark (51K multi-hop QA with bbox annotations), and significantly outperforms text-based…
…Generated by Qwen/Qwen2.5-Coder-32B-Instruct Vision Transformers (ViTs) have become a dominant architecture for visual representation learning, providing exceptionally strong and broadly reusable backbone features. However, ViTs are commonly…
…Generated by Qwen/Qwen2.5-Coder-32B-Instruct Reusable skills are a key mechanism for extending agent capabilities, allowing agents to accumulate experience and solve increasingly complex tasks. Yet most existing skill…
…a foveated Qwen adds +29 pts (~100%) of accuracy over the global baseline. Uniformly cranking resolution 10× buys almost nothing (+3.8 pts). Same lesson as the synthetic tasks: locality + recurrence beats…
…Generated by Qwen/Qwen2.5-Coder-32B-Instruct Existing image editing methods can be generally categorized into textual instruction-based and visual prompt-based ones. Textual instructions are semantically expressive, but are…
…Claude reached 97.7%, while GLM-5.2 still failed 12%, GPT-5.6 Sol 18%, and Qwen3-235B 42%. For some models, better localization traded incomplete deletion for over-editing. In…
…Generated by Qwen/Qwen2.5-Coder-32B-Instruct Recent advances in agentic Reinforcement Learning (RL) have substantially improved the multi-turn tool-use capabilities of large language model agents. However, most existing…
…Thanks for sharing!! Probably unrelated question, the graph shows Qwen3 1.7B has 2B parameters. Is it correct? · Model page show that's why we made this msitake But it's because…
…Generated by Qwen/Qwen2.5-Coder-32B-Instruct Although Large Language Model (LLM) agents have demonstrated strong performance on complex tasks, their learning is often limited by inefficient interaction feedback and static…