Researchers at the Gaoling School of Artificial Intelligence, Renmin University of China, have developed YuLan-Mini, a new open data-efficient language model featuring 2.42 billion parameters. This model is designed to enhance long-context capabilities and employs advanced training techniques, marking a notable advancement in the field of artificial intelligence. Additionally, the concept of Low-Rank Adaptation (LoRA) has emerged as a new approach to fine-tuning large language models, which allows for more efficient model training by significantly reducing the number of parameters that need to be updated during the fine-tuning process.
[LG] Cross-Entropy Is All You Need To Invert the Data Generating Process (2024) https://t.co/9x3uTiBmZF https://t.co/ocw4bkl4X3
[LG] RB-Modulation: Training-Free Personalization using Stochastic Optimal Control (2024) https://t.co/3mWWEwo2lC https://t.co/or4NgR7SEL
Inference-Aware Fine-Tuning for Best-of-N Sampling in Large Language Models https://t.co/Su8XEzTOPr