Abstract: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.
| Comments: | BEA Workshop 2026 |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2605.13624 [cs.CL] |
| (or arXiv:2605.13624v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2605.13624 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Takumi Goto [view email]
[v1]
Wed, 13 May 2026 14:52:15 UTC (146 KB)