EnrichIndex: Using LLMs to Enrich Retrieval Indices Offline

If you find our code, or the paper helpful, please cite the paper

@article{chen2025enrichindex,
  title={EnrichIndex: Using LLMs to Enrich Retrieval Indices Offline},
  author={Chen, Peter Baile and Wolfson, Tomer and Cafarella, Michael and Roth, Dan},
  journal={arXiv preprint arXiv:2504.03598},
  year={2025}
}

Overview

• enrichment.py includes the prompt for enriching documents/ tables
• This link includes the enrichment generated using gpt-4o-mini on
  • Reasoning-intensive retrieval: Bright
  • Complex table retrieval: Spider2, Beaver, Fiben

Contact

If you have any questions or feedback, please send an email to peterbc@mit.edu.