We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:


References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: Can Very Large Pretrained Language Models Learn Storytelling With A Few Examples?

Abstract: While pre-trained language models can generate individually fluent sentences for automatic story generation, they struggle to generate stories that are coherent, sensible and interesting. Current state-of-the-art (SOTA) story generation models explore using higher-level features such as plots or commonsense knowledge to improve the quality of generated stories. Prompt-based learning using very large pre-trained language models (VLPLMs) such as GPT3 has demonstrated impressive performance even across various NLP tasks. In this paper, we present an extensive study using automatic and human evaluation to compare the story generation capability of VLPLMs to those SOTA models in three different datasets where stories differ in style, register and length. Our results show that VLPLMs generate much higher quality stories than other story generation models, and to a certain extent rival human authors, although preliminary investigation also reveals that they tend to ``plagiarise'' real stories in scenarios that involve world knowledge.
Subjects: Computation and Language (cs.CL)
Cite as: arXiv:2301.09790 [cs.CL]
  (or arXiv:2301.09790v2 [cs.CL] for this version)

Submission history

From: Zhuohan Xie [view email]
[v1] Tue, 24 Jan 2023 02:44:02 GMT (7072kb,D)
[v2] Wed, 15 Mar 2023 23:47:00 GMT (7072kb,D)

Link back to: arXiv, form interface, contact.