The NLP community recently saw the release of a new large open-access multilingual language model, BLOOM (BigScience et al., 2022) covering 46 languages. We focus on BLOOM's multilingual ability by evaluating its machine translation performance across several datasets (WMT, Flores-101 and DiaBLa) and language pairs (high- and low-resourced). Our results show that 0-shot performance suffers from overgeneration and generating in the wrong language, but this is greatly improved in the few-shot setting, with very good results for a number of language pairs. We study several aspects including prompt design, model sizes, cross-lingual transfer and the use of discursive context.

翻译：自然语言处理领域近期发布了覆盖46种语言的大型开放多语言语言模型BLOOM（BigScience等，2022）。我们聚焦于BLOOM的多语言能力，通过评估其在多个数据集（WMT、Flores-101和DiaBLa）和语言对（高资源和低资源）上的机器翻译性能展开研究。结果表明，零样本设置下的翻译性能受过度生成及目标语言错误问题影响，但在少样本设置中这一问题得到显著改善，众多语言对取得了优异结果。我们研究了提示设计、模型规模、跨语言迁移及篇章语境使用等多个维度。