When2Think: Learning Difficulty-Aware Length Control for Efficient Hybrid Reasoning Models Paper • 2609.19671 • Published 10 days ago • 53
deepseek-ai/DeepSeek-V4-Flash-Vision-Exp Image-Text-to-Text • 305B • Updated 26 days ago • 925k • • 928
Knowing When Not to Reuse: Conditional Experience Transfer in Autonomous LLM Post-Training Paper • 2608.26730 • Published about 1 month ago • 155
LLaDA-Image: Building Strong Image Generators with Fully Open Training Recipes Paper • 2609.03796 • Published 24 days ago • 186
Training Chemical Plausibility-Aware Large Language Models for Single-Step Retrosynthesis Paper • 2608.18940 • Published Aug 19 • 35
OpenART: Scaling Agent Red Teaming via Open-Ended Environment Evolution Paper • 2608.00677 • Published Aug 1 • 265
MirrorWorld: Taming Video Diffusion Models for Mirror Reflection Generation Paper • 2608.07463 • Published Aug 7 • 6
Weights or Skills? A Survey of Robot-Learning Techniques: from Action-Predicting Weights to Robots that Write their Own Skills Paper • 2608.01851 • Published Aug 3 • 13
Learning from Failures: Retrieval-Centric CoT via Hard Negatives for Unified Multimodal Retrieval Paper • 2608.06060 • Published Aug 6 • 41