PyTerrier-GenRank: The PyTerrier Plugin for Reranking with Large Language Models

This provides a tool for researchers and practitioners in information retrieval to streamline LLM-based reranking experiments, but it is incremental as it builds on existing PyTerrier and LLM frameworks.

The authors tackled the challenge of experimenting with hyperparameters for using large language models as rerankers by introducing PyTerrier-GenRank, a PyTerrier plugin that facilitates seamless reranking experiments with support for strategies like pointwise and listwise prompting, validated through HuggingFace and OpenAI endpoints.