MiQA: A Benchmark for Inference on Metaphorical Questions
This work addresses the challenge of assessing metaphorical inference in AI, which is incremental as it builds on existing tasks but integrates them into a new benchmark.
The authors tackled the problem of evaluating large language models' ability to reason with metaphors by creating a benchmark that combines metaphor detection and commonsense reasoning, finding that performance ranges from chance to near-human levels depending on model size and prompting.
We propose a benchmark to assess the capability of large language models to reason with conventional metaphors. Our benchmark combines the previously isolated topics of metaphor detection and commonsense reasoning into a single task that requires a model to make inferences by accurately selecting between the literal and metaphorical register. We examine the performance of state-of-the-art pre-trained models on binary-choice tasks and find a large discrepancy between the performance of small and very large models, going from chance to near-human level. We also analyse the largest model in a generative setting and find that although human performance is approached, careful multiple-shot prompting is required.