IF-RewardBench: Benchmarking Judge Models for Instruction-Following Evaluation
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866917412667916288 |
|---|---|
| author | Wen, Bosi Niu, Yilin Wang, Cunxiang Ling, Xiaoying Zhang, Ying Ke, Pei Wang, Hongning Huang, Minlie |
| author_facet | Wen, Bosi Niu, Yilin Wang, Cunxiang Ling, Xiaoying Zhang, Ying Ke, Pei Wang, Hongning Huang, Minlie |
| contents | Instruction-following is a foundational capability of large language models (LLMs), with its improvement hinging on scalable and accurate feedback from judge models. However, the reliability of current judge models in instruction-following remains underexplored due to several deficiencies of existing meta-evaluation benchmarks, such as their insufficient data coverage and oversimplified pairwise evaluation paradigms that misalign with model optimization scenarios. To this end, we propose IF-RewardBench, a comprehensive meta-evaluation benchmark for instruction-following that covers diverse instruction and constraint types. For each instruction, we construct a preference graph containing all pairwise preferences among multiple responses based on instruction-following quality. This design enables a listwise evaluation paradigm that assesses the capabilities of judge models to rank multiple responses, which is essential in guiding model alignment. Extensive experiments on IF-RewardBench reveal significant deficiencies in current judge models and demonstrate that our benchmark achieves a stronger positive correlation with downstream task performance compared to existing benchmarks. Our codes and data are available at https://github.com/thu-coai/IF-RewardBench. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_04738 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | IF-RewardBench: Benchmarking Judge Models for Instruction-Following Evaluation Wen, Bosi Niu, Yilin Wang, Cunxiang Ling, Xiaoying Zhang, Ying Ke, Pei Wang, Hongning Huang, Minlie Computation and Language Instruction-following is a foundational capability of large language models (LLMs), with its improvement hinging on scalable and accurate feedback from judge models. However, the reliability of current judge models in instruction-following remains underexplored due to several deficiencies of existing meta-evaluation benchmarks, such as their insufficient data coverage and oversimplified pairwise evaluation paradigms that misalign with model optimization scenarios. To this end, we propose IF-RewardBench, a comprehensive meta-evaluation benchmark for instruction-following that covers diverse instruction and constraint types. For each instruction, we construct a preference graph containing all pairwise preferences among multiple responses based on instruction-following quality. This design enables a listwise evaluation paradigm that assesses the capabilities of judge models to rank multiple responses, which is essential in guiding model alignment. Extensive experiments on IF-RewardBench reveal significant deficiencies in current judge models and demonstrate that our benchmark achieves a stronger positive correlation with downstream task performance compared to existing benchmarks. Our codes and data are available at https://github.com/thu-coai/IF-RewardBench. |
| title | IF-RewardBench: Benchmarking Judge Models for Instruction-Following Evaluation |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2603.04738 |