Prompt Consistency for Zero-Shot Task Generalization

By Chunting Zhou, Junxian He, Xuezhe Ma, Taylor Berg-Kirkpatrick, Graham Neubig. 2022

97 citations recorded by the source corpus.

One of the most impressive results of recent NLP history is the ability of pre-trained language models to solve new tasks in a zero-shot setting. To achieve this, NLP tasks are framed as natural language prompts, generating a response indicating the predicted output. Nonetheless, the performance in such settings often lags far behind its supervised counterpart, suggesting a large space for potential improvement. In this paper, we explore methods to utilize unlabeled data to improve zero-shot performance. Specifically, we take advantage of the fact that multiple prompts can be used to specify a single task, and propose to regularize prompt consistency, encouraging consistent predictions over this diverse set of prompts. Our method makes it possible to fine-tune the model either with extra unlabeled training data, or directly on test input at inference time in an unsupervised manner. In experiments, our approach outperforms the state-of-the-art zero-shot learner, T0 (Sanh et al., 2022), on 9 out of 11 datasets across 4 NLP tasks by up to 10.6 absolute points in terms of accuracy. The gains are often attained with a small number of unlabeled examples.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Large Language Models, Prompting & Reasoning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Large Language Models

In Prompting & Reasoning

Citations in the corpus

SQL-PaLM: Improved Large Language Model Adaptation for Text-to-SQL (2024, unknown): “diverse of sampling for consistency [15, 16].”
Doubly Right Object Recognition: A Why Prompt for Visual Rationales (2022, methodology): “Prompt tuning is a lightweight adaptation method for language task [64–66].”
Discovering Latent Knowledge in Language Models Without Supervision (2022, background): “Particularly relevant are methods that also leverage unsupervised consistency properties, such as Jung et al. (2022); Zhou et al. (2022).”
One Embedder, Any Task: Instruction-Finetuned Text Embeddings (2022, background): “Previous work (Sanh et al., 2022; Zhou et al., 2022) shows that instruction-finetuned language models are not robust to paraphrased instructions.”
Test-Time Prompt Tuning for Zero-Shot Generalization in Vision-Language Models (2022, background): “Another line of work proposes to tune the prompt in an unsupervised manner [31, 32].”