Fine-grained Controllable Text Generation through In-context Learning with Feedback

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Thillainathan, Sarubi, Koller, Alexander
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866929388052807680
author Thillainathan, Sarubi
Koller, Alexander
author_facet Thillainathan, Sarubi
Koller, Alexander
contents We present a method for rewriting an input sentence to match specific values of nontrivial linguistic features, such as dependency depth. In contrast to earlier work, our method uses in-context learning rather than finetuning, making it applicable in use cases where data is sparse. We show that our model performs accurate rewrites and matches the state of the art on rewriting sentences to a specified school grade level.
format Preprint
id arxiv_https___arxiv_org_abs_2406_11338
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Fine-grained Controllable Text Generation through In-context Learning with Feedback
Thillainathan, Sarubi
Koller, Alexander
Computation and Language
We present a method for rewriting an input sentence to match specific values of nontrivial linguistic features, such as dependency depth. In contrast to earlier work, our method uses in-context learning rather than finetuning, making it applicable in use cases where data is sparse. We show that our model performs accurate rewrites and matches the state of the art on rewriting sentences to a specified school grade level.
title Fine-grained Controllable Text Generation through In-context Learning with Feedback
topic Computation and Language
url https://arxiv.org/abs/2406.11338