Post
REALIGN is a new method designed to improve the alignment of Large Language Models (LLMs) with human values by reformatting instruction data. This approach enhances LLM performance across various metrics by aligning responses with predefined criteria and evidence.
Key points:
* REALIGN has three steps: criteria definition, retrieval augmentation, and response reformatting
* It rewrites pairs (query, response) to enhance data quality for fine-tuning LLMs.
* It has shown significant improvements in general alignment, math reasoning and other tasks.
Congrats to the authors for their work!
Paper: Reformatted Alignment (2402.12219)
Code: https://github.com/GAIR-NLP/ReAlign
Key points:
* REALIGN has three steps: criteria definition, retrieval augmentation, and response reformatting
* It rewrites pairs (query, response) to enhance data quality for fine-tuning LLMs.
* It has shown significant improvements in general alignment, math reasoning and other tasks.
Congrats to the authors for their work!
Paper: Reformatted Alignment (2402.12219)
Code: https://github.com/GAIR-NLP/ReAlign
