Edit-level Majority Voting Mitigates Over-Correction in LLM-based Grammatical Error Correction
Fuente:
arXiv
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866910216595963904 |
|---|---|
| author | Goto, Takumi Sakai, Yusuke Watanabe, Taro |
| author_facet | Goto, Takumi Sakai, Yusuke Watanabe, Taro |
| contents | Grammatical error correction using large language models often suffers from the over-correction issue. To mitigate this, we propose a training-free inference method that performs edit-level majority voting over multiple candidates generated by a single model, without requiring model modifications or additional training. Across nine benchmarks covering English, Czech, German, Ukrainian, Korean, Hindi, and Romanian, the proposed method outperforms both greedy and MBR decoding in most cases. Moreover, it yields stable correction quality regardless of the instruction prompts used. We release two repository supporting GEC datasets loading and LLM inference. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_13624 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Edit-level Majority Voting Mitigates Over-Correction in LLM-based Grammatical Error Correction Goto, Takumi Sakai, Yusuke Watanabe, Taro Computation and Language Grammatical error correction using large language models often suffers from the over-correction issue. To mitigate this, we propose a training-free inference method that performs edit-level majority voting over multiple candidates generated by a single model, without requiring model modifications or additional training. Across nine benchmarks covering English, Czech, German, Ukrainian, Korean, Hindi, and Romanian, the proposed method outperforms both greedy and MBR decoding in most cases. Moreover, it yields stable correction quality regardless of the instruction prompts used. We release two repository supporting GEC datasets loading and LLM inference. |
| title | Edit-level Majority Voting Mitigates Over-Correction in LLM-based Grammatical Error Correction |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2605.13624 |