Fine-Tuning LLaMA for Effective Multi-Stage Text Retrieval, Controlling Evolving Toxicity in Language Models through Retrieval, and More!

Vol.21 for Oct 09 - Oct 15, 2023

Stay Ahead of the Curve with the Latest Advancements in Information Retrieval and Discoveries.

This week’s newsletter highlights the following research:

1. Policy Gradient Training for Effective Text Retrieval with Large Language Models, from Cornell

2. Improving Task-Specific Dense Retrieval via Adaptive Residual Embeddings, from Microsoft

3. Exploiting Long-term …