Toward Zero-Shot Instruction Following

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Lou, Renze, Yin, Wenpeng
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916105637855232
author Lou, Renze
Yin, Wenpeng
author_facet Lou, Renze
Yin, Wenpeng
contents This work proposes a challenging yet more realistic setting for zero-shot cross-task generalization: zero-shot instruction following, presuming the existence of a paragraph-style task definition while no demonstrations exist. To better learn the task supervision from the definition, we propose two strategies: first, to automatically find out the critical sentences in the definition; second, a ranking objective to force the model to generate the gold outputs with higher probabilities when those critical parts are highlighted in the definition. The joint efforts of the two strategies yield state-of-the-art performance on the Super-NaturalInstructions. Our code is available on GitHub.
format Preprint
id arxiv_https___arxiv_org_abs_2308_03795
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Toward Zero-Shot Instruction Following
Lou, Renze
Yin, Wenpeng
Computation and Language
This work proposes a challenging yet more realistic setting for zero-shot cross-task generalization: zero-shot instruction following, presuming the existence of a paragraph-style task definition while no demonstrations exist. To better learn the task supervision from the definition, we propose two strategies: first, to automatically find out the critical sentences in the definition; second, a ranking objective to force the model to generate the gold outputs with higher probabilities when those critical parts are highlighted in the definition. The joint efforts of the two strategies yield state-of-the-art performance on the Super-NaturalInstructions. Our code is available on GitHub.
title Toward Zero-Shot Instruction Following
topic Computation and Language
url https://arxiv.org/abs/2308.03795