Possible Stories: Evaluating Situated Commonsense Reasoning under Multiple Possible Scenarios
This addresses a gap in natural language processing for evaluating models on realistic, varied scenarios, though it is incremental as it builds on existing commonsense reasoning tasks.
The study tackled the problem of situated commonsense reasoning under multiple possible scenarios by creating a dataset called Possible Stories, which includes over 4.5K questions from 1.3K story texts, and found that current pretrained language models achieve only 60.2% accuracy in an unsupervised setting, far behind human accuracy of 92.5%.
The possible consequences for the same context may vary depending on the situation we refer to. However, current studies in natural language processing do not focus on situated commonsense reasoning under multiple possible scenarios. This study frames this task by asking multiple questions with the same set of possible endings as candidate answers, given a short story text. Our resulting dataset, Possible Stories, consists of more than 4.5K questions over 1.3K story texts in English. We discover that even current strong pretrained language models struggle to answer the questions consistently, highlighting that the highest accuracy in an unsupervised setting (60.2%) is far behind human accuracy (92.5%). Through a comparison with existing datasets, we observe that the questions in our dataset contain minimal annotation artifacts in the answer options. In addition, our dataset includes examples that require counterfactual reasoning, as well as those requiring readers' reactions and fictional information, suggesting that our dataset can serve as a challenging testbed for future studies on situated commonsense reasoning.