In-context: November 3, 2025

Written By Alyssa Pradhan
 

In-context: November 3, 2025

Here’s a quick wrap of the three papers we found interesting over the last few weeks with some take home points.

  • 00:35 - Scaling Large Language Models for Next-Generation Single-Cell Analysis

  • 05:55 - Generative Medical Event Models Improve with Scale

  • 10:50 - When helpfulness backfires: LLMs and the risk of false medical information due to sycophantic behavior 

Some resources and papers we discuss:

Rizvi, AS et al Scaling Large Language Models for Next-Generation Single-Cell Analysis, bioRxiv 2025.04.14.648850; doi: https://doi.org/10.1101/2025.04.14.648850

  • The Google blog post associated with the paper: https://blog.google/technology/ai/google-gemma-ai-cancer-therapy-discovery/

Waxler, Shane et al. Generative Medical Event Models Improve with Scale. ArXiv abs/2508.12104 (2025): n. pag. <https://arxiv.org/abs/2508.12104>

  • The ETHOS study referenced by this paper: Renc P, et al Zero shot health trajectory prediction using transformer. npj Digital Medicine, 7(1), Sep 2024. ISSN 2398-6352. doi: 10.1038/s41746-024-01235-0.

Chen S et al When helpfulness backfires: LLMs and the risk of false medical information due to sycophantic behavior. NPJ Digit Med. 2025 Oct 17;8(1):605. doi: 10.1038/s41746-025-02008-z. PMID: 41107408; PMCID: PMC12534679.

 
 
Alyssa Pradhan
Next

Should we be using LLMs for discharge summarisation?