LongStraw: Long-Context RL Beyond 2M Tokens under a Fixed GPU Budget Paper • 2607.14952 • Published 7 days ago • 188
MinT: Managed Infrastructure for Training and Serving Millions of LLMs Paper • 2605.13779 • Published May 13 • 225
$δ$-mem: Efficient Online Memory for Large Language Models Paper • 2605.12357 • Published May 12 • 132
$δ$-mem: Efficient Online Memory for Large Language Models Paper • 2605.12357 • Published May 12 • 132
SlowBA: An efficiency backdoor attack towards VLM-based GUI agents Paper • 2603.08316 • Published Mar 9 • 1
SlowBA: An efficiency backdoor attack towards VLM-based GUI agents Paper • 2603.08316 • Published Mar 9 • 1
PlanViz: Evaluating Planning-Oriented Image Generation and Editing for Computer-Use Tasks Paper • 2602.06663 • Published Feb 6 • 5
PlanViz: Evaluating Planning-Oriented Image Generation and Editing for Computer-Use Tasks Paper • 2602.06663 • Published Feb 6 • 5
AgentDevel: Reframing Self-Evolving LLM Agents as Release Engineering Paper • 2601.04620 • Published Jan 8 • 3
Golden Goose: A Simple Trick to Synthesize Unlimited RLVR Tasks from Unverifiable Internet Text Paper • 2601.22975 • Published Jan 30 • 113
view post Post 2246 From Pointers to Footnotes: Why Agents will be “Reference-First”https://medium.com/@di-zhang-fdu/from-pointers-to-footnotes-why-agents-will-be-reference-first-88d0e9730e37 See translation 👀 2 2 + Reply
view post Post 489 AgentDevel: Reframing Self-Evolving LLM Agents as Release EngineeringImprove your agent like Software Engineers, Driven by test and statistical information.https://www.arxiv.org/abs/2601.04620 See translation 👍 1 1 + Reply
Beyond Static Tools: Test-Time Tool Evolution for Scientific Reasoning Paper • 2601.07641 • Published Jan 12 • 48
AgentDevel: Reframing Self-Evolving LLM Agents as Release Engineering Paper • 2601.04620 • Published Jan 8 • 3
view post Post 343 On the eve of the birth of the Vibe Computer: Why agentic coding feels like a 1980s terminal — and why that’s the pointhttps://di-zhang-fdu.medium.com/on-the-eve-of-the-birth-of-the-vibe-computer-why-agentic-coding-feels-like-a-1980s-terminal-and-96c41510c356 See translation 2 replies · ❤️ 1 1 👀 1 1 + Reply
view post Post 472 EZTinker: A easy tinker-style RL-as-a-service demohttps://github.com/trotsky1997/EZTinker See translation 👀 2 2 + Reply
Error-Free Linear Attention is a Free Lunch: Exact Solution from Continuous-Time Dynamics Paper • 2512.12602 • Published Dec 14, 2025 • 44