Daily Papers of 2026-07-24

  1. AREX: Towards a Recursively Self-Improving Agent for Deep Research 149 upvotes, #1 of 2026-07-24
  2. K12-KGraph: A Curriculum-Aligned Knowledge Graph for Benchmarking and Training Educational LLMs 63 upvotes, #2 of 2026-07-24
  3. ReferTrack: Referring Then Tracking for Embodied Visual Tracking 53 upvotes, #3 of 2026-07-24
  4. Visual Contrastive Self-Distillation 51 upvotes, #4 of 2026-07-24
  5. SANA-Video 2.0: Hybrid Linear Attention with Attention Residuals for Efficient Video Generation 39 upvotes, #5 of 2026-07-24
  6. Show, Don't Tell: Evaluating Spatial Cognition in Generative Pixels Rather Than LLM Text 37 upvotes, #6 of 2026-07-24
  7. NVIDIA-labs OO Agents: Native Python Object-Oriented Agents 33 upvotes, #7 of 2026-07-24
  8. Tencent WorkBuddy Bench: A Multi-Domain Coding-Agent Benchmark with Contamination-Resistant Task Construction 26 upvotes, #8 of 2026-07-24
  9. LLMs Get Lost in Evolving User Intent 24 upvotes, #9 of 2026-07-24
  10. Color Pass-Through via Camera-Display Coupling 21 upvotes, #10 of 2026-07-24
  11. Sample-Efficient Learning from Agent Experience 19 upvotes, #11 of 2026-07-24
  12. Self-Supervised Learning of Structured Dynamics from Videos 19 upvotes, #11 of 2026-07-24
  13. Streaming Multi-Agent Autoregressive Diffusion Model with World State Registers 16 upvotes, #13 of 2026-07-24
  14. Multi-Turn On-Policy Distillation with Prefix Replay 12 upvotes, #14 of 2026-07-24
  15. Predictive Divergence Masks for LLM RL 11 upvotes, #15 of 2026-07-24
  16. Robostral Navigate 10 upvotes, #16 of 2026-07-24
  17. Recurrent Sinusoidal INRs for Efficient High-Fidelity Representation 9 upvotes, #17 of 2026-07-24
  18. FinanceComplexQA: Benchmarking Agentic Reasoning on Industrial-grade Financial Documents 8 upvotes, #18 of 2026-07-24
  19. OpenForgeRL: Train Harness-native Agents in Any Environment 8 upvotes, #18 of 2026-07-24
  20. GraphVid: Interactive Graph-Controllable Video Generation 8 upvotes, #18 of 2026-07-24
  21. TableVerse: A Large-scale Tabletop Dataset with Real-world Grounded Layouts for Generalizable Manipulation 7 upvotes, #21 of 2026-07-24
  22. Dataset Distillation by Influence Matching 3 upvotes, #22 of 2026-07-24

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.