Won't Get Fooled Again: Answering Questions with False Premises

TL;DR

This paper investigates how pre-trained language models can be activated to effectively rebut false premise questions, revealing their inherent knowledge and proposing methods to improve their question-answering capabilities.

Contribution

The study introduces the FalseQA dataset and demonstrates that fine-tuning PLMs enables them to discriminate false premise questions and generate reasonable rebuttals.

Findings

PLMs can distinguish false premise questions after moderate fine-tuning.

PLMs generate plausible explanations for false premises.

Replay training with general questions enhances PLMs' performance on FPQs and general QA.

Abstract

Pre-trained language models (PLMs) have shown unprecedented potential in various fields, especially as the backbones for question-answering (QA) systems. However, they tend to be easily deceived by tricky questions such as "How many eyes does the sun have?". Such frailties of PLMs often allude to the lack of knowledge within them. In this paper, we find that the PLMs already possess the knowledge required to rebut such questions, and the key is how to activate the knowledge. To systematize this observation, we investigate the PLMs' responses to one kind of tricky questions, i.e., the false premises questions (FPQs). We annotate a FalseQA dataset containing 2365 human-written FPQs, with the corresponding explanations for the false premises and the revised true premise questions. Using FalseQA, we discover that PLMs are capable of discriminating FPQs by fine-tuning on moderate numbers (e.g., 256) of examples. PLMs also generate reasonable explanations for the false premise, which serve as rebuttals. Further replaying a few general questions during training allows PLMs to excel on FPQs and general questions simultaneously. Our work suggests that once the rebuttal ability is stimulated, knowledge inside the PLMs can be effectively utilized to handle FPQs, which incentivizes the research on PLM-based QA systems.