Grammar Concept Annotation at Scale: Deployed Fine-Tuned Small Language Models Outperform Prompted Frontier Models

查看原文 / View Original 编辑 / Edit

Computer Science > Computation and Language arXiv:2610.10827 (cs) [Submitted on 7 Oct 2026] Title: Grammar Concept Annotation at Scale: Deployed Fine-Tuned Small Language Models Outperform Prompted Frontier Models Authors: Marjan Celikik, Ana Peleteiro Ramallo, Javier Morales.

计算机科学 > 计算与语言 arXiv:2610.10827 (cs) [提交于 2026 年 10 月 7 日] 标题：大规模语法概念标注：部署的微调小型语言模型优于提示的前沿模型作者：Marjan Celikik, Ana Peleteiro Ramallo, Javier Morales。

Abstract: Corrective feedback is among the best-evidenced drivers of second-language acquisition, yet corrections delivered during lessons rarely accumulate into an actionable view of grammar mastery. Prompted frontier models can provide such a view from learner—tutor lesson transcripts, but they are costly at scale.

摘要：纠正性反馈是第二语言习得中证据最充分的驱动因素之一，然而在课程中提供的纠正很少能积累成对语法掌握程度的可操作性视图。提示的前沿模型可以从学习者与导师的课程记录中提供此类视图，但它们在大规模应用时成本高昂。

We close this gap by fine-tuning Qwen3.5 small language models (SLMs) on filtered and rebalanced teacher-generated supervision, then deploying an efficient 0.8B model in an end-to-end grammar mastery tracker for all English learners on our platform.

我们通过在经过筛选和重新平衡的教师生成监督数据上微调 Qwen3.5 小型语言模型 (SLM) 来弥补这一差距，随后在我们的平台上为所有英语学习者部署了一个高效的 0.8B 模型，用于端到端的语法掌握跟踪。

Internalizing the annotation contract into adapter weights enables pairing the 0.8B model with a compact matched prompt rather than verbose instructions. On two human-curated benchmarks, both the deployed 0.8B model and a 4B reference comparator outperform prompted GPT-5.4 and GPT-5.6 Sol in precision and recall under nested matching criteria of increasing strictness: concept, evidence span, and correctness.

将标注契约内化到适配器权重中，使得 0.8B 模型能够与简洁的匹配提示配对，而非冗长的指令。在两个人工策划的基准测试中，部署的 0.8B 模型和 4B 参考对比模型在概念、证据范围和正确性等日益严格的嵌套匹配标准下，其精确率和召回率均优于提示的 GPT-5.4 和 GPT-5.6 Sol。

The deployed 0.8B SLM reduces serving cost by approximately 16$\times$. A feature-level online experiment shows significant gains in learner engagement ($+15.8%$) and key business metrics, including scheduled hours ($+2.1%$) and GMV from new lessons ($+13.2%$).

部署的 0.8B SLM 将服务成本降低了约 16 倍。一项功能级的在线实验显示，学习者参与度（+15.8%）以及关键业务指标（包括预定课时 +2.1% 和新课程带来的商品交易总额 GMV +13.2%）均有显著增长。