Llama 3.2 summaries of bioRxiv and medRxiv preprints with biorecap

The biorecap package can now summarize bioRxiv and medRxiv preprints, and uses the llama-3.2-3b model by default.

Last month I published a paper and an R package for summarizing preprints from bioRxiv using a local LLM. I wrote about it here:

biorecap: an R package for summarizing bioRxiv preprints with a local LLM

Llama 3.2 was just released today (announcement). The biggest news is the addition of a multimodal vision model, but I was intrigued by the reasonably good performance of the tiny 3B text model.

I used this as an excuse to update the biorecap R package. The default model is now llama3.2-3b and it’s noticeably faster, but the larger update is that biorecap now supports medRxiv in addition to bioRxiv. The documentation describes how this works, but essentially you give it any number of subject areas from bioRxiv and/or medRxiv, and it figures out which RSS feed to pull recent papers from before summarizing with llama3.2-3b by default.

As an example, I’m pulling bioinformatics articles from bioRxiv and infectious diseases articles from medRxiv, summarizing with llama3.2-3b¹:

biorecap_report(subject = c("bioinformatics", "infectious_diseases"),
                model = "llama3.2")

Results from biorecap_report() run on September 25, 2024. Each section is an excerpt from a larger report with 30 articles for each subject.

See the package documentation for details.

1. You need to pull the model first: ollama pull llama3.2.