Method

LLM reranking

Using an LLM to rerank candidates, pointwise or listwise.

16 cases from 15 companies apply LLM reranking. 2 were checked in an online A/B test; 2 report results in production.

Typical tasks: Rank search results, Retrieve and rank from millions of items, Build an assistant over internal docs, Match entities and 6 more. Companies: Meta, Databricks, Sber, Target, DoorDash, Instacart and 9 more. Often combined with: Embedding similarity matching, LLM-as-a-judge, Dense retrieval, Hybrid search, RAG and 41 more.