Daily Papers of 2026-07-15

  1. Function-Aware Fill-in-the-Middle as Mid-Training for Coding Agent Foundation Models 107 upvotes, #1 of 2026-07-15
  2. Search Beyond What Can Be Taught: Evolving the Knowledge Boundary in Agentic Visual Generation 86 upvotes, #2 of 2026-07-15
  3. Read It Back: Pretrained MLLMs Are Zero-Shot Reward Models for Text-to-Image Generation 83 upvotes, #3 of 2026-07-15
  4. SynthDocBench: Controlled Benchmark for Long-Context Visual Document Understanding 70 upvotes, #4 of 2026-07-15
  5. Blind-Spots-Bench: Evaluating Blind Spots in Multimodal Models 35 upvotes, #5 of 2026-07-15
  6. Know Before Fix: QA-Driven Repository Knowledge Acquisition for Software Issue Resolution 22 upvotes, #6 of 2026-07-15
  7. MuScriptor: An Open Model for Multi-Instrument Music Transcription 21 upvotes, #7 of 2026-07-15
  8. Towards Autonomous and Auditable Medical Imaging Model Development 20 upvotes, #8 of 2026-07-15
  9. Navigating the Mirage: A Dual-Path Agentic Framework for Robust Misleading Chart Question Answering 17 upvotes, #9 of 2026-07-15
  10. MonkeyOCRv2: A Visual-Text Foundation Model for Document AI 16 upvotes, #10 of 2026-07-15
  11. What LLM Forecasters Know but Don't Say: Probing Internal Representations for Calibration and Faithfulness 14 upvotes, #11 of 2026-07-15
  12. Let RGB Be the Language of Vision 14 upvotes, #11 of 2026-07-15
  13. Principled Analysis of Deep Reinforcement Learning Evaluation and Design Paradigms 10 upvotes, #13 of 2026-07-15
  14. Are LLMs Ready for Scientific Discovery? A Capability-Oriented Benchmark for AI Scientists 10 upvotes, #13 of 2026-07-15
  15. MAGIC: Transition-Aware Generation of Navigable Multi-Scene Game Worlds with Large Language Models 7 upvotes, #15 of 2026-07-15

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.