Daily Papers of 2025-06-19

  1. Sekai: A Video Dataset towards World Exploration 60 upvotes, #1 of 2025-06-19
  2. GenRecal: Generation after Recalibration from Large to Small Vision-Language Models 36 upvotes, #2 of 2025-06-19
  3. All is Not Lost: LLM Recovery without Checkpoints 35 upvotes, #3 of 2025-06-19
  4. ProtoReasoning: Prototypes as the Foundation for Generalizable Reasoning in LLMs 31 upvotes, #4 of 2025-06-19
  5. Embodied Web Agents: Bridging Physical-Digital Realms for Integrated Agent Intelligence 23 upvotes, #5 of 2025-06-19
  6. BUT System for the MLC-SLM Challenge 14 upvotes, #6 of 2025-06-19
  7. Semantically-Aware Rewards for Open-Ended R1 Training in Free-Form Generation 12 upvotes, #7 of 2025-06-19
  8. SciVer: Evaluating Foundation Models for Multimodal Scientific Claim Verification 11 upvotes, #8 of 2025-06-19
  9. SwarmAgentic: Towards Fully Automated Agentic System Generation via Swarm Intelligence 11 upvotes, #8 of 2025-06-19
  10. ImmerseGen: Agent-Guided Immersive World Generation with Alpha-Textured Proxies 10 upvotes, #10 of 2025-06-19
  11. Truncated Proximal Policy Optimization 10 upvotes, #10 of 2025-06-19
  12. MoTE: Mixture of Ternary Experts for Memory-efficient Large Multimodal Models 8 upvotes, #12 of 2025-06-19
  13. CoMemo: LVLMs Need Image Context with Image Memory 7 upvotes, #13 of 2025-06-19
  14. GMT: General Motion Tracking for Humanoid Whole-Body Control 7 upvotes, #13 of 2025-06-19
  15. FedNano: Toward Lightweight Federated Tuning for Pretrained Multimodal Large Language Models 7 upvotes, #13 of 2025-06-19
  16. PictSure: Pretraining Embeddings Matters for In-Context Learning Image Classifiers 7 upvotes, #13 of 2025-06-19
  17. AssertBench: A Benchmark for Evaluating Self-Assertion in Large Language Models 4 upvotes, #17 of 2025-06-19
  18. OS-Harm: A Benchmark for Measuring Safety of Computer Use Agents 4 upvotes, #17 of 2025-06-19
  19. Evolutionary Caching to Accelerate Your Off-the-Shelf Diffusion Model 4 upvotes, #17 of 2025-06-19

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.