Answer, assemble, ace: Understanding how transformers answer multiple choice questions

S Wiegreffe, O Tafjord, Y Belinkov, H Hajishirzi… - arXiv preprint arXiv …, 2024 - arxiv.org
Multiple-choice question answering (MCQA) is a key competence of performant transformer
language models that is tested by mainstream benchmarks. However, recent evidence …

A Study on Large Language Models' Limitations in Multiple-Choice Question Answering

A Khatun, DG Brown - arXiv preprint arXiv:2401.07955, 2024 - arxiv.org
The widespread adoption of Large Language Models (LLMs) has become commonplace,
particularly with the emergence of open-source models. More importantly, smaller models …

How Chinese are Chinese Language Models? The Puzzling Lack of Language Policy in China's LLMs

AW Wen-Yi, UES Jo, LJ Lin, D Mimno - arXiv preprint arXiv:2407.09652, 2024 - arxiv.org
Contemporary language models are increasingly multilingual, but Chinese LLM developers
must navigate complex political and business considerations of language diversity …