Factual Probing Is [MASK]: Learning vs. Learning to Recall

Abstract

Petroni et al. (2019) demonstrated that it is possible to retrieve world facts from a pre-trained language model by expressing them as cloze-style prompts and interpret the model's prediction accuracy as a lower bound on the amount of factual information it encodes. Subsequent work has attempted to tighten the estimate by searching for better prompts, using a disjoint set of facts as training data. In this work, we make two complementary contributions to better understand these factual probing techniques. First, we propose OptiPrompt, a novel and efficient method which directly optimizes in continuous embedding space. We find this simple method is able to predict an additional 6.4% of facts in the LAMA benchmark. Second, we raise a more important question: Can we really interpret these probing results as a lower bound? Is it possible that these prompt-search methods learn from the training data too? We find, somewhat surprisingly, that the training data used by these methods contains certain regularities of the underlying fact distribution, and all the existing prompt methods, including ours, are able to exploit them for better fact prediction. We conduct a set of control experiments to disentangle "learning" from "learning to recall", providing a more detailed picture of what different prompts can reveal about pre-trained language models.

NAACL 2021. The code is publicly available at https://github.com/princeton-nlp/OptiPrompt

References in corpus (4)

• Language Models are Few-Shot Learners
• Prefix-Tuning: Optimizing Continuous Prompts for Generation
• Right for the Wrong Reasons: Diagnosing Syntactic Heuristics in Natural Language Inference
• An information theoretic view on selecting linguistic probes

Cited by in corpus (7)

• P-Tuning v2: Prompt Tuning Can Be Comparable to Fine-tuning Universally Across Scales and Tasks
• True Few-Shot Learning with Language Models
• Recent Advances in Natural Language Processing via Large Pre-Trained Language Models: A Survey
• Paradigm Shift in Natural Language Processing
• Differentiable Prompt Makes Pre-trained Language Models Better Few-shot Learners
• An Empirical Study on Few-shot Knowledge Probing for Pretrained Language Models
• Can Language Models be Biomedical Knowledge Bases?