# Re-autoresearching MSMARCO BM25, on Vespa

DevFeed: [Re-autoresearching MSMARCO BM25, on Vespa](<https://devfeed.tech/articles/re-autoresearching-msmarco-bm25-on-vespa-12796.md>)

Original publisher: [Read original article](<https://blog.vespa.ai/re-autoresearching-msmarco-bm25-on-vespa/>)

Author: andreer thomas

Published: 2026-05-29T00:00:00Z

Content type: article

Language: en

Sources: [Vespa Blog](<https://devfeed.tech/sources/vespa-blog.md>)

Topics: [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [dataset](<https://devfeed.tech/topics/dataset.md>), [Python](<https://devfeed.tech/topics/python.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [generalization in machine learning](<https://devfeed.tech/topics/generalization-in-machine-learning.md>), [pandas](<https://devfeed.tech/topics/pandas.md>), [Google Search](<https://devfeed.tech/topics/google-search.md>)

Tags: [benchmark](<https://devfeed.tech/tags/benchmark.md>), [bm25](<https://devfeed.tech/tags/bm25.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [experiment](<https://devfeed.tech/tags/experiment.md>), [google-search](<https://devfeed.tech/tags/google-search.md>), [information-retrieval](<https://devfeed.tech/tags/information-retrieval.md>), [openai](<https://devfeed.tech/tags/openai.md>), [pandas](<https://devfeed.tech/tags/pandas.md>), [python](<https://devfeed.tech/tags/python.md>), [ranking](<https://devfeed.tech/tags/ranking.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [zero-shot](<https://devfeed.tech/tags/zero-shot.md>)

## AI overview

This article reproduces an MSMARCO BM25 autoresearch experiment in Vespa. It compares LLM-driven Python reranking with an approach restricted to existing Vespa rank features and reports a comparable improvement on a 650,000-passage subset, with better generalization to the full dataset.

## Source excerpt

BM25 is having a moment. We reproduce Doug Turnbull's MSMARCO autoresearch experiment in Vespa and get a comparable MRR@10 lift from existing rank features -- with twice the generalization to full MSMARCO.