Spaces:
Running
Running
| title: README | |
| emoji: 馃悪 | |
| colorFrom: blue | |
| colorTo: indigo | |
| sdk: static | |
| pinned: true | |
| <div align="center"> | |
| <h1>SightAndCite</h1> | |
| <p><b>Retrieval that sees the structure and cites the evidence.</b></p> | |
| <p>Open research on retrieval-augmented generation (RAG), chunking, multi-hop reasoning and efficient LLM systems.</p> | |
| </div> | |
| --- | |
| ## 馃敪 Featured: SpectRAG | |
| **Stop retrieving chunks. Start retrieving context.** | |
| SpectRAG replaces independent top-k ranking with **context-set selection** over a weighted chunk graph and its Laplacian spectrum. It returns evidence that is relevant, diverse and connected, all under a fixed token budget. | |
| <a href="https://huggingface.co/spaces/sightandcite/SpectRAG"> | |
| <img src="https://huggingface.co/spaces/sightandcite/SpectRAG/resolve/main/assets/01_hero_topk_vs_spectrag.gif" alt="Classic top-k retrieval finds 1 of 3 gold passages; SpectRAG finds 3 of 3" width="100%"> | |
| </a> | |
| | | | | |
| |---|---| | |
| | 馃搫 **Paper** | [arXiv:2610.36991](https://arxiv.org/abs/2610.36991) 路 馃 [Paper page](https://huggingface.co/papers/2610.36991) | | |
| | 馃幀 **Visual walkthrough** | [sightandcite/SpectRAG](https://huggingface.co/spaces/sightandcite/SpectRAG) | | |
| | 馃搳 **Evaluation results** | [sightandcite/SpectRAG-eval-results](https://huggingface.co/datasets/sightandcite/SpectRAG-eval-results) | | |
| | 馃捇 **Code (MIT)** | [github.com/SightAndCite/SpectRAG](https://github.com/SightAndCite/SpectRAG) | | |
| **Highlights on MultiHopRAG & UltraDomain** | |
| - **91.8%** overall gold-evidence recall (next best 76.1%) | |
| - **95.1%** exact match on inference questions | |
| - **~6脳 smaller index** than LightRAG (46 MB vs 274 MB) | |
| - LLM judges prefer SpectRAG's answers for **diversity** against every baseline | |
| --- | |
| ## 馃Л What we work on | |
| - **Retrieval-augmented generation:** context-set selection, hybrid retrieval, graph-based RAG | |
| - **Chunking:** structure-aware and semantic segmentation for long documents | |
| - **Multi-hop systems:** recovering bridge evidence across chunks and documents | |
| - **Efficient LLM pipelines:** distillation, lightweight indexes, fully local deployments | |
| ## 馃懃 Team | |
| Zafer Sahinoglu 路 Abdulazeez Mohammed 路 Buse Yaren Kazangirler 路 Cahit Berkay Kazangirler | |
| ## 馃摤 Get in touch | |
| Questions, ideas or a retriever that breaks in interesting ways? Open a discussion on the [paper page](https://huggingface.co/papers/2610.36991) or an issue on [GitHub](https://github.com/SightAndCite/SpectRAG/issues). |