REPAIR Scientific Retrievers
LoRA adapters (on Qwen2.5) for REPAIR: Resolving Long-Tail Confusion in Scientific Retrievers via Fact-Verified Iterative Refinement. Code: https://github.com/yerimoh/REPAIR
| Size | Base | Repository | Avg. (paper) |
|---|---|---|---|
| 0.5B | Qwen/Qwen2.5-0.5B | yerim0210/REPAIR_Scientific_Retrievers_0.5B | 0.583 |
| 1.5B | Qwen/Qwen2.5-1.5B | yerim0210/REPAIR_Scientific_Retrievers_1.5B | 0.607 |
| 7B | Qwen/Qwen2.5-7B | yerim0210/REPAIR_Scientific_Retrievers_7B | 0.623 |
Citation
@misc{oh2026repairresolvinglongtailconfusion,
title={REPAIR: Resolving Long-Tail Confusion in Scientific Retrievers via Fact-Verified Iterative Refinement},
author={Yerim Oh and Gunhee Kim},
year={2026},
booktitle={Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing},
}