---
license: mit
task_categories:
- text-retrieval
tags:
- scientific-papers
- retrieval
- benchmark
---

# SciFullBench

This dataset corresponds to the paper [Chain of Retrieval: Multi-Aspect Iterative Search Expansion and Post-Order Search Aggregation for Full Paper Retrieval](https://huggingface.co/papers/2507.10057). It provides a large-scale benchmark with complete and segmented contexts of full scientific papers for document-to-document retrieval.

The official code and evaluation pipeline are available at: [https://github.com/psw0021/Chain-of-Retrieval-Official](https://github.com/psw0021/Chain-of-Retrieval-Official)

**License:** MIT