For Generated Text, Is NLI-Neutral Text the Best Text?
Fuente:
arXiv
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2023
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913201434656768 |
|---|---|
| author | Mersinias, Michail Mahowald, Kyle |
| author_facet | Mersinias, Michail Mahowald, Kyle |
| contents | We explore incorporating natural language inference (NLI) into the text generative pipeline by using a pre-trained NLI model to assess whether a generated sentence entails, contradicts, or is neutral to the prompt and preceding text. First, we show that the NLI task is predictive of generation errors made by GPT-3. We use these results to develop an NLI-informed generation procedure for GPT-J. Then, we evaluate these generations by obtaining human annotations on error types and overall quality. We find that an NLI strategy of maximizing entailment improves text generation when the nucleus sampling randomness parameter value is high, while one which maximizes contradiction is in fact productive when the parameter value is low. Overall, though, we demonstrate that an NLI strategy of maximizing the neutral class provides the highest quality of generated text (significantly better than the vanilla generations), regardless of parameter value. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2302_08577 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | For Generated Text, Is NLI-Neutral Text the Best Text? Mersinias, Michail Mahowald, Kyle Computation and Language We explore incorporating natural language inference (NLI) into the text generative pipeline by using a pre-trained NLI model to assess whether a generated sentence entails, contradicts, or is neutral to the prompt and preceding text. First, we show that the NLI task is predictive of generation errors made by GPT-3. We use these results to develop an NLI-informed generation procedure for GPT-J. Then, we evaluate these generations by obtaining human annotations on error types and overall quality. We find that an NLI strategy of maximizing entailment improves text generation when the nucleus sampling randomness parameter value is high, while one which maximizes contradiction is in fact productive when the parameter value is low. Overall, though, we demonstrate that an NLI strategy of maximizing the neutral class provides the highest quality of generated text (significantly better than the vanilla generations), regardless of parameter value. |
| title | For Generated Text, Is NLI-Neutral Text the Best Text? |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2302.08577 |