Daily Papers of 2025-02-04

  1. OmniHuman-1: Rethinking the Scaling-Up of One-Stage Conditioned Human Animation Models 168 upvotes, #1 of 2025-02-04
  2. The Differences Between Direct Alignment Algorithms are a Blur 109 upvotes, #2 of 2025-02-04
  3. Process Reinforcement through Implicit Rewards 53 upvotes, #3 of 2025-02-04
  4. Preference Leakage: A Contamination Problem in LLM-as-a-judge 34 upvotes, #4 of 2025-02-04
  5. AlignVLM: Bridging Vision and Language Latent Spaces for Multimodal Understanding 33 upvotes, #5 of 2025-02-04
  6. SafeRAG: Benchmarking Security in Retrieval-Augmented Generation of Large Language Model 25 upvotes, #6 of 2025-02-04
  7. SliderSpace: Decomposing the Visual Capabilities of Diffusion Models 24 upvotes, #7 of 2025-02-04
  8. MM-IQ: Benchmarking Human-Like Abstraction and Reasoning in Multimodal Models 22 upvotes, #8 of 2025-02-04
  9. DeepRAG: Thinking to Retrieval Step by Step for Large Language Models 21 upvotes, #9 of 2025-02-04
  10. MakeAnything: Harnessing Diffusion Transformers for Multi-Domain Procedural Sequence Generation 20 upvotes, #10 of 2025-02-04
  11. Scaling Embedding Layers in Language Models 20 upvotes, #10 of 2025-02-04
  12. AIN: The Arabic INclusive Large Multimodal Model 15 upvotes, #12 of 2025-02-04
  13. FastKV: KV Cache Compression for Fast Long-Context Processing with Token-Selective Propagation 14 upvotes, #13 of 2025-02-04
  14. ZebraLogic: On the Scaling Limits of LLMs for Logical Reasoning 14 upvotes, #13 of 2025-02-04
  15. The Jumping Reasoning Curve? Tracking the Evolution of Reasoning Performance in GPT-[n] and o-[n] Models on Multimodal Puzzles 12 upvotes, #15 of 2025-02-04
  16. Almost Surely Safe Alignment of Large Language Models at Inference-Time 11 upvotes, #16 of 2025-02-04
  17. RandLoRA: Full-rank parameter-efficient fine-tuning of large models 9 upvotes, #17 of 2025-02-04
  18. PhD Knowledge Not Required: A Reasoning Challenge for Large Language Models 9 upvotes, #17 of 2025-02-04
  19. Improving Transformer World Models for Data-Efficient RL 9 upvotes, #17 of 2025-02-04
  20. Improved Training Technique for Latent Consistency Models 7 upvotes, #20 of 2025-02-04
  21. Learning to Generate Unit Tests for Automated Debugging 4 upvotes, #21 of 2025-02-04
  22. Lifelong Sequential Knowledge Editing without Model Degradation 4 upvotes, #21 of 2025-02-04
  23. LongDPO: Unlock Better Long-form Generation Abilities for LLMs via Critique-augmented Stepwise Information 4 upvotes, #21 of 2025-02-04
  24. A Study on the Performance of U-Net Modifications in Retroperitoneal Tumor Segmentation 3 upvotes, #24 of 2025-02-04
  25. Current Pathology Foundation Models are unrobust to Medical Center Differences 2 upvotes, #25 of 2025-02-04
  26. Language Models Prefer What They Know: Relative Confidence Estimation via Confidence Preferences 2 upvotes, #25 of 2025-02-04

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.