Inductive Linguistic Reasoning with Large Language Models
This work addresses the need to assess and enhance LLMs' abstract reasoning skills for low-resource languages, which is incremental as it builds on existing prompting methods.
The paper tackled the problem of evaluating large language models' linguistic reasoning capabilities by using analogical prompting to improve performance on multilingual puzzles, boosting GPT-4o by up to 8.1% and Llama-3.1-405B-Instruct by 5.9% over chain-of-thought approaches.
Evaluating large language models (LLMs) on their linguistic reasoning capabilities is an important task to understand the gaps in their skills that may surface during large-scale adoption. In this work, we investigate the abilities of such models to perform abstract multilingual reasoning through the lens of linguistic puzzles on extremely low-resource languages. As these translation tasks involve inductive and deductive reasoning from reference instances, we examine whether diverse auxiliary demonstrations can be automatically induced from seed exemplars, through analogical prompting. We employ a two-stage procedure, first generating analogical exemplars with a language model, and then applying them in-context along with provided target language exemplars. Our results on the modeLing dataset show that analogical prompting is effective in eliciting models' knowledge of language grammar similarities, boosting the performance of GPT-4o by as much as 8.1% and Llama-3.1-405B-Instruct by 5.9% over chain-of-thought approaches. These gains are attributable to the analogical demonstrations, both when self-generated as well as when produced by weaker multilingual models. Furthermore, we demonstrate that our method generalizes to other tasks present in Linguistics Olympiad competitions, achieving sizable improvements across all problem types and difficulty levels included in the LINGOLY dataset with GPT-4o. We also report several findings about interesting phenomena which drive linguistic reasoning performance, suggesting that such puzzles are a valuable benchmark for new reasoning methods.