Daily Papers of 2025-08-06

  1. Seed Diffusion: A Large-Scale Diffusion Language Model with High-Speed Inference 120 upvotes, #1 of 2025-08-06
  2. Skywork UniPic: Unified Autoregressive Modeling for Visual Understanding and Generation 58 upvotes, #2 of 2025-08-06
  3. LongVie: Multimodal-Guided Controllable Ultra-Long Video Generation 49 upvotes, #3 of 2025-08-06
  4. CompassVerifier: A Unified and Robust Verifier for LLMs Evaluation and Outcome Reward 32 upvotes, #4 of 2025-08-06
  5. Tool-integrated Reinforcement Learning for Repo Deep Search 18 upvotes, #5 of 2025-08-06
  6. Representation Shift: Unifying Token Compression with FlashAttention 15 upvotes, #6 of 2025-08-06
  7. LiveMCPBench: Can Agents Navigate an Ocean of MCP Tools? 13 upvotes, #7 of 2025-08-06
  8. CRINN: Contrastive Reinforcement Learning for Approximate Nearest Neighbor Search 13 upvotes, #7 of 2025-08-06
  9. The Promise of RL for Autoregressive Image Editing 11 upvotes, #9 of 2025-08-06
  10. Goedel-Prover-V2: Scaling Formal Theorem Proving with Scaffolded Data Synthesis and Self-Correction 10 upvotes, #10 of 2025-08-06
  11. LAMIC: Layout-Aware Multi-Image Composition via Scalability of Multimodal Diffusion Transformer 8 upvotes, #11 of 2025-08-06
  12. Multi-human Interactive Talking Dataset 8 upvotes, #11 of 2025-08-06
  13. HyCodePolicy: Hybrid Language Controllers for Multimodal Monitoring and Decision in Embodied Agents 5 upvotes, #13 of 2025-08-06
  14. ChartCap: Mitigating Hallucination of Dense Chart Captioning 5 upvotes, #13 of 2025-08-06
  15. UniEgoMotion: A Unified Model for Egocentric Motion Reconstruction, Forecasting, and Generation 4 upvotes, #15 of 2025-08-06
  16. What Is Your AI Agent Buying? Evaluation, Implications and Emerging Questions for Agentic E-Commerce 4 upvotes, #15 of 2025-08-06
  17. Bidirectional Likelihood Estimation with Multi-Modal Large Language Models for Text-Video Retrieval 3 upvotes, #17 of 2025-08-06
  18. AlignGuard-LoRA: Alignment-Preserving Fine-Tuning via Fisher-Guided Decomposition and Riemannian-Geodesic Collision Regularization 2 upvotes, #18 of 2025-08-06
  19. AttnTrace: Attention-based Context Traceback for Long-Context LLMs 2 upvotes, #18 of 2025-08-06
  20. TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs 1 upvotes, #20 of 2025-08-06
  21. TreeRanker: Fast and Model-agnostic Ranking System for Code Suggestions in IDEs 1 upvotes, #20 of 2025-08-06

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.