What do Large Language Models Learn about Scripts?

TL;DR

This paper investigates whether large language models inherently understand script knowledge by testing their ability to generate event sequences from scenarios, and proposes a pipeline to improve their performance in this task.

Contribution

The paper introduces a novel task of generating event sequence descriptions from scenarios and proposes a pipeline-based framework (SIF) to enhance LMs' script knowledge understanding.

Findings

Zero-shot LMs produce poor event sequences.

SIF significantly improves event sequence quality.

Manual analysis indicates further research is needed.

Abstract

Script Knowledge (Schank and Abelson, 1975) has long been recognized as crucial for language understanding as it can help in filling in unstated information in a narrative. However, such knowledge is expensive to produce manually and difficult to induce from text due to reporting bias (Gordon and Van Durme, 2013). In this work, we are interested in the scientific question of whether explicit script knowledge is present and accessible through pre-trained generative language models (LMs). To this end, we introduce the task of generating full event sequence descriptions (ESDs) given a scenario in the form of natural language prompts. In zero-shot probing experiments, we find that generative LMs produce poor ESDs with mostly omitted, irrelevant, repeated or misordered events. To address this, we propose a pipeline-based script induction framework (SIF) which can generate good quality ESDs for unseen scenarios (e.g., bake a cake). SIF is a two-staged framework that fine-tunes LM on a small set of ESD examples in the first stage. In the second stage, ESD generated for an unseen scenario is post-processed using RoBERTa-based models to filter irrelevant events, remove repetitions, and reorder the temporally misordered events. Through automatic and manual evaluations, we demonstrate that SIF yields substantial improvements ($1$-$3$ BLUE points) over a fine-tuned LM. However, manual analysis shows that there is great room for improvement, offering a new research direction for inducing script knowledge.