appliedml42
@appliedml42
abs: https://arxiv.org/abs/2302.00083v1 TLDR: This paper proposes an alternative approach to Retrieval-Augmented Language Modeling (RALM) called In-Context RALM, which prepends relevant documents from a grounding corpus to the input of an unchanged language model. https://i.imgur.com/JBFZWcN.png
1 reply
0 recast
0 reaction
appliedml42
@appliedml42
keywords: Retrieval-Augmented Language Modeling | Language Model | LM Architecture | Document Retrieval and Ranking | LM Grounding
0 reply
0 recast
0 reaction
