TY - CPAPER TI - Language Models Are Poor Learners of Directional Inference AU - Li, Tianyi AU - Hosseini, Mohammad Javad AU - Weber, Sabine AU - Steedman, Mark PY - 2022 T2 - Findings of the Association for Computational Linguistics: EMNLP 2022 SP - 903–921 PB - Association for Computational Linguistics DO - 10.18653/v1/2022.findings-emnlp.64 UR - https://fis.uni-bamberg.de/handle/uniba/97222 LA - eng AB - We examine LMs’ competence of directional predicate entailments by supervised fine-tuning with prompts. Our analysis shows that contrary to their apparent success on standard NLI, LMs show limited ability to learn such directional inference; moreover, existing datasets fail to test directionality, and/or are infested by artefacts that can be learnt as proxy for entailments, yielding over-optimistic results. In response, we present BoOQA (Boolean Open QA), a robust multi-lingual evaluation benchmark for directional predicate entailments, extrinsic to existing training sets. On BoOQA, we establish baselines and show evidence of existing LM-prompting models being incompetent directional entailment learners, in contrast to entailment graphs, however limited by sparsity. KW - Language Models ER -