Tests large language models on humanities datasets, revealing model performance variability and cost implications.
The RISE Humanities Data Benchmark replaces anecdotal model selection with systematic evidence by testing a wide range of large language models across diverse humanities datasets, revealing that no single model dominates, cost differences between similarly performing models can be dramatic, and that the process of making evaluative standards explicit is itself a key benefit for research projects.
No takes yet. Share an insight, caveat, or question.
Hindermann et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: