arXiv：cs.CL· Yirui Zhang, Yash Raj Shrestha·· 16 小时前
研究：LLM 审查很少能胜过更强模型，需区分验证与能力迁移
Review Rarely Beats the Stronger Model: Separating Verification from Capability Transfer in LLM Pipelines
AI 导读
一项针对 12 个开源模型（3B 至 72B）48 对组合的研究发现，在数学、知识、代码和推理任务中，审查仅在 96 个审查方向中的 9 个胜过更强的模型。Qwen3-8B 审查 Llama-3.1-8B 虽提升 33.2 分，却比 Qwen3-8B 单独作答低 6.4 分，原因在于审查者受锚定效应影响多跟随生成者、丢弃自身正确答案。
正文
Abstract:Qwen3-8B reviewing Llama-3.1-8B in math adds 33.2 points, yet Qwen3-8B alone is 6.4 points more accurate than the result. Reviewing another model's answer is usually judged by its gain over the generator. We show that this confuses verification, which adds accuracy beyond either model alone, with capability transfer, where a stronger reviewer supplies answers it could have given on its own. Measured against the stronger model alone, across 48 pairs of 12 open-weight models (3B to 72B) in math, knowledge, code and reasoning, review beats it in only 9 of 96 review directions, and does worse the larger the capability gap. The reason is anchoring: the reviewer mostly follows the generator, discarding its own correct answers. Review costs 2.6 to 2.8 times as much as the stronger model alone, while a selector using the same two calls is 2.7 to 4.2 points more accurate than review and cheaper. Studies of verification should report the stronger model alone, not only the gain over the generator.
| Comments: | 22 pages, 7 figures, 12 tables. Code: this https URL |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.15897 [cs.CL] |
| (or arXiv:2610.15897v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.15897 arXiv-issued DOI via DataCite |
Submission history
From: Yirui Zhang [view email]
[v1]
Sun, 11 Oct 2026 20:26:36 UTC (288 KB)
[v2]
Wed, 14 Oct 2026 11:38:52 UTC (291 KB)
来源：arXiv：cs.CL · arxiv.org