SEAL: Can Saturated Benchmarks Be Revived by LLM-as-a-Meta-Judge?
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913170694602752 |
|---|---|
| author | Chen, Jiamin Wu, Yidi Wang, Qiexiang Chen, Qianben Li, Yuchen Zhang, Yansen Zhang, Xiaokun Zhou, Wangchunshu Ma, Chen |
| author_facet | Chen, Jiamin Wu, Yidi Wang, Qiexiang Chen, Qianben Li, Yuchen Zhang, Yansen Zhang, Xiaokun Zhou, Wangchunshu Ma, Chen |
| contents | Widely used language-model benchmarks are increasingly saturated, with frontier systems often receiving near-tied scores that standard metrics cannot resolve. Rather than constructing harder alternatives, we ask whether existing tasks can be made informative again through improved evaluation over the same candidate outputs. Therefore, we present Seeded Elimination with Adaptive LLM-as-a-Meta-Judge, a self-improving evaluation protocol for extracting latent ranking signal from saturated benchmarks. SEAL seeds candidate outputs into a single elimination and evaluates each match with task-level principles plus self-improving checklist criteria. We evaluate SEAL on multiple saturated benchmarks covering code generation, mathematical reasoning, knowledge-intensive question answering, and tool-use agent task completion. Across these settings, SEAL improves the ranking-accuracy--latency trade-off over competing protocols, attaining 0.83--1.00 Spearman agreement with full pairwise judging and 4/4 top-1 agreement, while requiring only 11.89 calls per task compared with 28.00 for full pairwise evaluation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_30104 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | SEAL: Can Saturated Benchmarks Be Revived by LLM-as-a-Meta-Judge? Chen, Jiamin Wu, Yidi Wang, Qiexiang Chen, Qianben Li, Yuchen Zhang, Yansen Zhang, Xiaokun Zhou, Wangchunshu Ma, Chen Computation and Language Widely used language-model benchmarks are increasingly saturated, with frontier systems often receiving near-tied scores that standard metrics cannot resolve. Rather than constructing harder alternatives, we ask whether existing tasks can be made informative again through improved evaluation over the same candidate outputs. Therefore, we present Seeded Elimination with Adaptive LLM-as-a-Meta-Judge, a self-improving evaluation protocol for extracting latent ranking signal from saturated benchmarks. SEAL seeds candidate outputs into a single elimination and evaluates each match with task-level principles plus self-improving checklist criteria. We evaluate SEAL on multiple saturated benchmarks covering code generation, mathematical reasoning, knowledge-intensive question answering, and tool-use agent task completion. Across these settings, SEAL improves the ranking-accuracy--latency trade-off over competing protocols, attaining 0.83--1.00 Spearman agreement with full pairwise judging and 4/4 top-1 agreement, while requiring only 11.89 calls per task compared with 28.00 for full pairwise evaluation. |
| title | SEAL: Can Saturated Benchmarks Be Revived by LLM-as-a-Meta-Judge? |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2605.30104 |