Trendoday
Hugging FaceArchive · daily

HF Daily Papers on October 28, 2026

On October 28, 2026, Trendoday checked HF Daily Papers 12 times and recorded 68 different entries. “ChronoReplay: One Answer Interface Is Not Enough for…” ranked highest over the day: on the list in 92% of checks, best #1, average #1.4. “Hy-Nomas: Towards Free-form Interactive World Generation” and “Rollback: Experience Backtracking in World Models for…” followed. 0 entries were on the list at every check. 43 entries made their first appearance, led by “ChronoReplay: One Answer Interface Is Not Enough for…”.

  1. 1
    92%of checks
  2. 2
    92%of checks
  3. 3
    92%of checks
  4. 4
    92%of checks
  5. 5
    92%of checks
  6. 6
    92%of checks
  7. 7
    75%of checks
  8. 8
    92%of checks
  9. 9
    75%of checks
  10. 10
    83%of checks
  11. 11
    OvisOCR3 Technical Report↗
    best #6 · avg #12.8new
    92%of checks
  12. 12
    92%of checks
  13. 13
    92%of checks
  14. 14
    75%of checks
  15. 15
    92%of checks
  16. 16
    92%of checks
  17. 17
    92%of checks
  18. 18
    67%of checks
  19. 19
    67%of checks
  20. 20
    75%of checks
  21. 21
    92%of checks
  22. 22
    67%of checks
  23. 23
    92%of checks
  24. 24
    83%of checks
  25. 25
    50%of checks
  26. 26
    75%of checks
  27. 27
    67%of checks
  28. 28
    75%of checks
  29. 29
    42%of checks
  30. 30
    50%of checks
  31. 31
    58%of checks
  32. 32
    33%of checks
  33. 33
    42%of checks
  34. 34
    42%of checks
  35. 35
    33%of checks
  36. 36
    42%of checks
  37. 37
    33%of checks
  38. 38
    25%of checks
  39. 39
    17%of checks
  40. 40
    17%of checks
  41. 41
    17%of checks
  42. 42
    8%of checks
  43. 43
    8%of checks
  44. 44
    8%of checks
  45. 45
    8%of checks
  46. 46
    8%of checks
  47. 47
    17%of checks
  48. 48
    8%of checks
  49. 49
    8%of checks
  50. 50
    17%of checks

Also on the list that day (18)

miniTE: Diagnosing and Bridging the Execution Gap in Native NVFP4… · S3AC: Stabilized, Selective, and Scalable Actor-Critic for LLM… · Build, Don't Play: Self-Improving LLM Research Agents Develop a… · EmbodiedLab: Benchmarking Embodied Agents That Act to Know · CausalAlign: Causal Alignment of Video Generation Models for… · Rethinking Video Storage for AI Training: Breaking Cross-Frame… · RynnVLA-Latent: Scaling Robotic Manipulation Foundation Models with… · AgentProofBench: Benchmarking Reusable Verifiers for Web Agents · More Memory, Better Videos? A Controlled Study of Visual Memory in… · Thinking-in-Bets at Test Time Reduces Sycophantic Revisions in LLM… · Novel or Not? PatEx-Gym: Agentic Prior-Art Search for Novelty… · Nested Sparsity for Scalable Pretraining and Continual Learning · Constrained Decoding: Harmful or Beneficial? Revisiting Its Impact… · Mycelium: A Meta-Study of Hallucination Detection Across Provenance,… · Trust Mechanisms for Offline Generative AI at the Edge: A… · Ψ-0.6: Adaptive Streams of Thought in a Unified Vision-Language… · Joint or Per-Option? Compact Encoders for Closed-Set Decisions from… · EnvLunch: The Environment as a Free Lunch for Learning to Interact

Ordered by average position weighted by how often the entry was on the list that day. Ranks are UTC days. All HF Daily Papers archives

Part of week 44 and October 2026.