Paper Image

Large language models generate knowledge for AI tasks

Published on:

11 October 2023

Primary Category:

Computation and Language

Paper Authors:

Liang Chen,

Yang Deng,

Yatao Bian,

Zeyu Qin,

Bingzhe Wu,

Tat-Seng Chua,

Kam-Fai Wong


Key Details

Large language models surpass retrieval for knowledge-intensive AI tasks

Models generate more relevant, coherent knowledge than retrieval

Generated knowledge has lower factuality than retrieved knowledge

Factuality issues don't always harm downstream task performance

Proposed evaluation framework assesses knowledge quality comprehensively

AI generated summary

Large language models generate knowledge for AI tasks

This paper evaluates large language models on their ability to generate knowledge for AI tasks like question answering and dialogue. The models outperform retrieval methods but raise concerns about factuality. A new framework is proposed to evaluate generated knowledge from multiple perspectives.

Answers from this paper


No comments yet, be the first to start the conversation...

Sign up to comment on this paper

Sign Up