Red teaming, the practice of adversarially exposing unexpected or undesired model behaviors, is critical towards improving equity and accuracy of large language models, but non-model creator-affiliated red teaming is scant in healthcare. We convened teams of clinicians, medical and engineering students, and technical professionals (80 participants total) to stress-test models with real-world clinical cases and categorize inappropriate responses along axes of safety, privacy, hallucinations/accuracy, and bias. Six medically-trained reviewers re-analyzed prompt-response pairs and added qualitative annotations. Of 376 unique prompts (1504 responses), 20.1% were inappropriate (GPT-3.5: 25.8%; GPT-4.0: 16%; GPT-4.0 with Internet: 17.8%). Subsequently, we show the utility of our benchmark by testing GPT-4o, a model released after our event (20.4% inappropriate). 21.5% of responses appropriate with GPT-3.5 were inappropriate in updated models. We share insights for constructing red teaming prompts, and present our benchmark for iterative model assessments.

The formose reaction (FR) autocatalytically converts simple plausibly prebiotic feedstocks into molecules of biological interest, including ribose. Autocatalysis is a hallmark of life, thus various studies have explored the formose reaction with respect to the origins of life. The FR is robust under appropriate conditions, occurring readily at low temperatures from various substrates, and has been implicated in the generation of meteoritic organic compounds. We explored the FR here using a combination of in silico modeling techniques and high resolution mass spectrometry. The models match experimental results well, and point to the FR being much more complex than previously modeled or measured, and help explain the FR’s potential to generate homochirality and primitive compartments, both of which are also hallmarks of life, before the emergence of the complex directed molecular encoding suggested by the RNA World model. These results suggest the FR requires further study with regard to the origins of life, and its importance may lie in the way it enables and coordinates emergent chemistries, rather than the particular products it generates, such as ribose.