Daily Papers of 2025-11-12

  1. Tiny Model, Big Logic: Diversity-Driven Optimization Elicits Large-Model Reasoning Ability in VibeThinker-1.5B 108 upvotes, #1 of 2025-11-12
  2. Adaptive Multi-Agent Response Refinement in Conversational Systems 39 upvotes, #2 of 2025-11-12
  3. KLASS: KL-Guided Fast Inference in Masked Diffusion Models 35 upvotes, #3 of 2025-11-12
  4. Beyond English: Toward Inclusive and Scalable Multilingual Machine Translation with LLMs 31 upvotes, #4 of 2025-11-12
  5. Wasm: A Pipeline for Constructing Structured Arabic Interleaved Multimodal Corpora 31 upvotes, #4 of 2025-11-12
  6. The Path Not Taken: RLVR Provably Learns Off the Principals 27 upvotes, #6 of 2025-11-12
  7. VideoSSR: Video Self-Supervised Reinforcement Learning 21 upvotes, #7 of 2025-11-12
  8. Beyond Fact Retrieval: Episodic Memory for RAG with Generative Semantic Workspaces 8 upvotes, #8 of 2025-11-12
  9. Intelligence per Watt: Measuring Intelligence Efficiency of Local AI 6 upvotes, #9 of 2025-11-12
  10. DynaAct: Large Language Model Reasoning with Dynamic Action Spaces 5 upvotes, #10 of 2025-11-12
  11. Optimizing Diversity and Quality through Base-Aligned Model Collaboration 4 upvotes, #11 of 2025-11-12
  12. Walking the Tightrope of LLMs for Software Development: A Practitioners' Perspective 4 upvotes, #11 of 2025-11-12
  13. TimeSearch-R: Adaptive Temporal Search for Long-Form Video Understanding via Self-Verification Reinforcement Learning 2 upvotes, #13 of 2025-11-12
  14. BiCA: Effective Biomedical Dense Retrieval with Citation-Aware Hard Negatives 2 upvotes, #13 of 2025-11-12

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.