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

Download:

Current browse context:

cs.CL

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

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

Computer Science > Computation and Language

Title: All That's 'Human' Is Not Gold: Evaluating Human Evaluation of Generated Text

Abstract: Human evaluations are typically considered the gold standard in natural language generation, but as models' fluency improves, how well can evaluators detect and judge machine-generated text? We run a study assessing non-experts' ability to distinguish between human- and machine-authored text (GPT2 and GPT3) in three domains (stories, news articles, and recipes). We find that, without training, evaluators distinguished between GPT3- and human-authored text at random chance level. We explore three approaches for quickly training evaluators to better identify GPT3-authored text (detailed instructions, annotated examples, and paired examples) and find that while evaluators' accuracy improved up to 55%, it did not significantly improve across the three domains. Given the inconsistent results across text domains and the often contradictory reasons evaluators gave for their judgments, we examine the role untrained human evaluations play in NLG evaluation and provide recommendations to NLG researchers for improving human evaluations of text generated from state-of-the-art models.
Comments: references added, corrected typo
Subjects: Computation and Language (cs.CL)
Cite as: arXiv:2107.00061 [cs.CL]
  (or arXiv:2107.00061v2 [cs.CL] for this version)

Submission history

From: Elizabeth Clark [view email]
[v1] Wed, 30 Jun 2021 19:00:25 GMT (7453kb,D)
[v2] Wed, 7 Jul 2021 06:07:22 GMT (7453kb,D)

Link back to: arXiv, form interface, contact.