DeepSeek-R1 vs. o3-mini: How Well can Reasoning LLMs Evaluate MT and Summarization?
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866916768538165248 |
|---|---|
| author | Larionov, Daniil Takeshita, Sotaro Zhang, Ran Chen, Yanran Leiter, Christoph Wang, Zhipin Greisinger, Christian Eger, Steffen |
| author_facet | Larionov, Daniil Takeshita, Sotaro Zhang, Ran Chen, Yanran Leiter, Christoph Wang, Zhipin Greisinger, Christian Eger, Steffen |
| contents | Reasoning-enabled large language models (LLMs) excel in logical tasks, yet their utility for evaluating natural language generation remains unexplored. This study systematically compares reasoning LLMs with non-reasoning counterparts across machine translation and text summarization evaluation tasks. We evaluate eight models spanning state-of-the-art reasoning models (DeepSeek-R1, OpenAI o3), their distilled variants (8B-70B parameters), and equivalent non-reasoning LLMs. Experiments on WMT23 and SummEval benchmarks reveal architecture and task-dependent benefits: OpenAI o3-mini models show improved performance with increased reasoning on MT, while DeepSeek-R1 and generally underperforms compared to its non-reasoning variant except in summarization consistency evaluation. Correlation analysis demonstrates that reasoning token usage correlates with evaluation quality only in specific models, while almost all models generally allocate more reasoning tokens when identifying more quality issues. Distillation maintains reasonable performance up to 32B parameter models but degrades substantially at 8B scale. This work provides the first assessment of reasoning LLMs for NLG evaluation and comparison to non-reasoning models. We share our code to facilitate further research: https://github.com/NL2G/reasoning-eval. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_08120 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DeepSeek-R1 vs. o3-mini: How Well can Reasoning LLMs Evaluate MT and Summarization? Larionov, Daniil Takeshita, Sotaro Zhang, Ran Chen, Yanran Leiter, Christoph Wang, Zhipin Greisinger, Christian Eger, Steffen Computation and Language Reasoning-enabled large language models (LLMs) excel in logical tasks, yet their utility for evaluating natural language generation remains unexplored. This study systematically compares reasoning LLMs with non-reasoning counterparts across machine translation and text summarization evaluation tasks. We evaluate eight models spanning state-of-the-art reasoning models (DeepSeek-R1, OpenAI o3), their distilled variants (8B-70B parameters), and equivalent non-reasoning LLMs. Experiments on WMT23 and SummEval benchmarks reveal architecture and task-dependent benefits: OpenAI o3-mini models show improved performance with increased reasoning on MT, while DeepSeek-R1 and generally underperforms compared to its non-reasoning variant except in summarization consistency evaluation. Correlation analysis demonstrates that reasoning token usage correlates with evaluation quality only in specific models, while almost all models generally allocate more reasoning tokens when identifying more quality issues. Distillation maintains reasonable performance up to 32B parameter models but degrades substantially at 8B scale. This work provides the first assessment of reasoning LLMs for NLG evaluation and comparison to non-reasoning models. We share our code to facilitate further research: https://github.com/NL2G/reasoning-eval. |
| title | DeepSeek-R1 vs. o3-mini: How Well can Reasoning LLMs Evaluate MT and Summarization? |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2504.08120 |