2023-04-24-ren23a.md

title	abstract	layout	series	publisher	issn	id	month	tex_title	firstpage	lastpage	page	order	cycles	bibtex_author	author	date	address	container-title	volume	genre	issued	pdf	extras
Self-Evaluation Improves Selective Generation in Large Language Models	Safe deployment of large language models (LLMs) may benefit from a reliable method for assessing their generated content to determine when to abstain or to selectively generate. While likelihood-based metrics such as perplexity are widely employed, recent research has demonstrated the limitations of using sequence-level probability estimates given by LLMs as reliable indicators of generation quality. Conversely, LLMs have demonstrated strong calibration at the token level, particularly when it comes to choosing correct answers in multiple-choice questions or evaluating true/false statements. In this work, we reformulate open-ended generation tasks into token-level prediction tasks, and leverage LLMs’ superior calibration at the token level. We instruct an LLM to self-evaluate its answers, employing either a multi-way comparison or a point-wise evaluation approach, with the option to include an “None of the above” option to express the model’s uncertainty explicitly. We benchmark a range of scoring methods based on self-evaluation and evaluate their performance in selective generation using TruthfulQA and TL;DR. Through extensive experiments with PaLM-2 and GPT-3, we demonstrate that self-evaluation based scores not only improve accuracy, but also correlate better with the overall quality of generated content.	inproceedings	Proceedings of Machine Learning Research	PMLR	2640-3498	ren23a	0	Self-Evaluation Improves Selective Generation in Large Language Models	49	64	49-64	49	false	Ren, Jie and Zhao, Yao and Vu, Tu and Liu, Peter J. and Lakshminarayanan, Balaji	given family Jie Ren given family Yao Zhao given family Tu Vu given family Peter J. Liu given family Balaji Lakshminarayanan	2023-04-24		Proceedings on "I Can't Believe It's Not Better: Failure Modes in the Age of Foundation Models" at NeurIPS 2023 Workshops	239	inproceedings	date-parts 2023 4 24	https://proceedings.mlr.press/v239/ren23a/ren23a.pdf