Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2503.06765 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916647958216704 |
|---|---|
| author | Takayama, Shusuke Frank, Ian |
| author_facet | Takayama, Shusuke Frank, Ian |
| contents | We compare the effectiveness of zero-shot Chain-of-Thought (CoT) prompting in Japanese and English using ChatGPT-3.5 and 4o-mini. The technique of zero-shot CoT, which involves appending a phrase such as "Let's think step by step" to a prompt to encourage reasoning before answering, has been shown to offer LLM performance improvements in mathematical and reasoning tasks, particularly in English. We investigate how these effects transfer to Japanese using the Japanese Multi-task Language Understanding Benchmark (JMMLU) and the Multi-task Language Understanding Benchmark (MMLU). Our results show that while zero-shot CoT prompting can lead to notable performance gains for some prompt categories in GPT-3.5, its impact in GPT-4o-mini is associated with significant performance declines. However, for Japanese prompts there remain certain categories, such as college mathematics and abstract algebra, that still exhibit improvements, despite the broader trend of diminishing effectiveness in more advanced models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_06765 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Effectiveness of Zero-shot-CoT in Japanese Prompts Takayama, Shusuke Frank, Ian Computation and Language Artificial Intelligence We compare the effectiveness of zero-shot Chain-of-Thought (CoT) prompting in Japanese and English using ChatGPT-3.5 and 4o-mini. The technique of zero-shot CoT, which involves appending a phrase such as "Let's think step by step" to a prompt to encourage reasoning before answering, has been shown to offer LLM performance improvements in mathematical and reasoning tasks, particularly in English. We investigate how these effects transfer to Japanese using the Japanese Multi-task Language Understanding Benchmark (JMMLU) and the Multi-task Language Understanding Benchmark (MMLU). Our results show that while zero-shot CoT prompting can lead to notable performance gains for some prompt categories in GPT-3.5, its impact in GPT-4o-mini is associated with significant performance declines. However, for Japanese prompts there remain certain categories, such as college mathematics and abstract algebra, that still exhibit improvements, despite the broader trend of diminishing effectiveness in more advanced models. |
| title | Effectiveness of Zero-shot-CoT in Japanese Prompts |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2503.06765 |