AI 日报 · 2026-07-23

Full-parameter post-training of trillion-parameter-scale MoE models introduces substantial system-level challenges for l…

HF Daily Papers · 2026-07-23

LLM agent failures are difficult to debug because the step where an error surfaces is often not the one that caused it. …

HF Daily Papers · 2026-07-22

Structural fidelity is essential to scientific methodology diagrams. To communicate research logic, these diagrams must …

HF Daily Papers · 2026-07-22

Asynchronous reinforcement learning improves throughput by decoupling rollout generation from optimization, but stalenes…

HF Daily Papers · 2026-07-22

Recent autoregressive video diffusion methods are increasingly built upon Self Forcing, where the student is trained on …

HF Daily Papers · 2026-07-23

Unlike conventional video game development, which relies on labor-intensive pipelines for asset production, animation, p…

HF Daily Papers · 2026-07-22

Large-scale visual generators are increasingly capable but costly to train, fine-tune, and deploy. We introduce Mage-Flo…

HF Daily Papers · 2026-07-22

Scaling executable agent training data for LLM post-training is bottlenecked by substrate-bound methods that tie task ge…

HF Daily Papers · 2026-07-21

We present ABot-World-0, an action-conditioned video world model for real-time, long-horizon closed-loop interaction, su…

HF Daily Papers · 2026-07-22

Large language models (LLMs) are increasingly used to automate data-processing workflows, yet coding agents typically pr…

HF Daily Papers · 2026-07-22

Mobile graphical user interface (GUI) agents have demonstrated remarkable capabilities in automating complex tasks, yet …

HF Daily Papers · 2026-07-21

As large language models and AI agents become the primary consumers of search results, document set quality determines t…

HF Daily Papers · 2026-07-23