On the Use of LLMs for Specialised Terminology: A Good Alternative to Corpora?
Quick Answer
This study evaluates LLMs like Claude Sonnet 4.5 and DeepSeek for specialized translation in Earth Sciences and NLP, finding that while LLMs can assist translators, they cannot fully replace specialized corpora.
Quick Take
Claude Sonnet 4.5 performed best under optimal conditions, while DeepSeek showed greater stability.
Key Points
- Evaluated four : GPT-4o, GPT-5.2, Claude Sonnet 4.5, and DeepSeek.
- Conducted experiments on 80 terms in Earth Sciences and NLP domains.
- Claude Sonnet 4.5 achieved the best results in favorable configurations.
- DeepSeek demonstrated greater stability across different prompting strategies.
- LLMs are useful but cannot replace specialized corpora at this stage.
DeepSignal Analysis
What happened
This study evaluates the effectiveness of large language models (LLMs) like Claude Sonnet 4.5 and DeepSeek in specialized translation for Earth Sciences and NLP. The research indicates that while LLMs can assist translators, they do not replace specialized corpora. Claude Sonnet 4.5 performed best under optimal conditions, while DeepSeek demonstrated greater stability.
Key evidence
- The study assessed four proprietary models: GPT-4o, GPT-5.2, Claude Sonnet 4.5, and DeepSeek, across two specialized domains: Earth, Environmental and Planetary Sciences, and Natural Language Processing.
- The experiment involved 80 terms per domain and compared two prompting strategies: terminology mode and translation mode, revealing differences in model performance.
- Claude Sonnet 4.5 achieved the best results in favorable conditions, while DeepSeek was noted for its stability, indicating varied performance across models.
Why it matters
The findings highlight the potential of LLMs as tools for specialized translation, suggesting they can aid translators but cannot fully substitute for traditional resources like specialized corpora. This distinction is crucial for understanding the current limitations of LLMs in professional and educational contexts.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Specialised translation relies on the use of documentary and terminological resources, including corpora. These resources are particularly useful for terminology. However, their compilation and exploitation have several limitations: they require time, technical skills and access to data that can be difficult to collect. This study examines the extent to which LLMs can assist specialised translators in finding equivalents from English to French. We evaluate four proprietary models, GPT-4o, GPT-5.2, Claude Sonnet 4.5 and DeepSeek, in two specialised domains, Earth, Environmental and Planetary Sciences (EEPS) and Natural Language Processing (NLP). The experiment is based on 80 terms per domain and compares two prompting strategies: a terminology and a translation mode. The results highlight clear differences between models, prompting strategies and, to a lesser extent, domains. Claude Sonnet 4.5 achieves the best results in the most favourable configuration, while DeepSeek stands out for its greater stability. Analysis of confidence estimates also shows that they are only a partial indicator of terminological accuracy. Overall, the findings suggest that LLMs can be useful tools for specialised translators, but cannot, at this stage, replace specialised corpora. This research therefore paves the way for future work on the real practical usefulness of LLMs for specialised translators in work and educational contexts.
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2607.24784 [cs.AI] |
| (or arXiv:2607.24784v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2607.24784 arXiv-issued DOI via DataCite |
|
| Journal reference: | 26th Annual Conference of the European Association for Machine Translation, Jun 2026, Tilburg, Netherlands |
Submission history
From: Joachim Minder [view email] [via CCSD proxy]
[v1]
Mon, 22 Jun 2026 09:03:13 UTC (931 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.AI
See more →HOBA: Hierarchical On-Policy Bidding Agents for Adaptive Online Advertising
HOBA (Hierarchical On-policy Bidding Agents) is a novel hierarchical reinforcement learning framework that enhances online advertising bidding systems by improving adaptability and reducing hyperparameter tuning costs. It utilizes a for hyperparameter inference, a SARSA agent for expert model selection, and a dynamic expert pool for bid execution, achieving a +3.6% increase in target cost during large-scale deployment and outperforming state-of-the-art baselines on AuctionNet.