modeLing: A Novel Dataset for Testing Linguistic Reasoning in Language Models
This provides a new benchmark for testing linguistic reasoning in AI systems, addressing data leakage issues, but it is incremental as it builds on existing evaluation frameworks.
The authors tackled the problem of evaluating linguistic reasoning in language models by introducing modeLing, a novel benchmark of Linguistics Olympiad-style puzzles, and found that large language models achieved non-negligible accuracy, demonstrating emergent few-shot reasoning abilities.
We introduce modeLing, a novel benchmark of Linguistics Olympiad-style puzzles which tests few-shot reasoning in AI systems. Solving these puzzles necessitates inferring aspects of a language's grammatical structure from a small number of examples. Such puzzles provide a natural testbed for language models, as they require compositional generalization and few-shot inductive reasoning. Consisting solely of new puzzles written specifically for this work, modeLing has no risk of appearing in the training data of existing AI systems: this ameliorates the risk of data leakage, a potential confounder for many prior evaluations of reasoning. Evaluating several large open source language models and GPT on our benchmark, we observe non-negligible accuracy, demonstrating few-shot emergent reasoning ability which cannot merely be attributed to shallow memorization. However, imperfect model performance suggests that modeLing can be used to measure further progress in linguistic reasoning.