Daily Papers of 2023-09-19

  1. CulturaX: A Cleaned, Enormous, and Multilingual Dataset for Large Language Models in 167 Languages 87 upvotes, #1 of 2023-09-19
  2. Adapting Large Language Models via Reading Comprehension 82 upvotes, #2 of 2023-09-19
  3. PDFTriage: Question Answering over Long, Structured Documents 55 upvotes, #3 of 2023-09-19
  4. Contrastive Decoding Improves Reasoning in Large Language Models 39 upvotes, #4 of 2023-09-19
  5. Sorted LLaMA: Unlocking the Potential of Intermediate Layers of Large Language Models for Dynamic Inference Using Sorted Fine-Tuning (SoFT) 23 upvotes, #5 of 2023-09-19
  6. An Empirical Study of Scaling Instruct-Tuned Large Multimodal Models 19 upvotes, #6 of 2023-09-19
  7. LayoutNUWA: Revealing the Hidden Layout Expertise of Large Language Models 15 upvotes, #7 of 2023-09-19
  8. Cure the headache of Transformers via Collinear Constrained Attention 13 upvotes, #8 of 2023-09-19
  9. MindAgent: Emergent Gaming Interaction 12 upvotes, #9 of 2023-09-19
  10. Struc-Bench: Are Large Language Models Really Good at Generating Complex Structured Data? 11 upvotes, #10 of 2023-09-19
  11. TextBind: Multi-turn Interleaved Multimodal Instruction-following 7 upvotes, #11 of 2023-09-19
  12. A Distributed Data-Parallel PyTorch Implementation of the Distributed Shampoo Optimizer for Training Neural Networks At-Scale 6 upvotes, #12 of 2023-09-19
  13. Recovering from Privacy-Preserving Masking with Large Language Models 4 upvotes, #13 of 2023-09-19
  14. Stack-and-Delay: a new codebook pattern for music generation 4 upvotes, #13 of 2023-09-19
  15. S3-DST: Structured Open-Domain Dialogue Segmentation and State Tracking in the Era of LLMs 4 upvotes, #13 of 2023-09-19
  16. Enhance audio generation controllability through representation similarity regularization 3 upvotes, #16 of 2023-09-19
  17. Augmenting text for spoken language understanding with Large Language Models 2 upvotes, #17 of 2023-09-19

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.