Back to feed
arXiv cs.AI·

Perovskite-R1: a domain-specialized large language model for intelligent discovery of precursor additives and experimental design

Signal
78
Hype
25
In three linesPerovskite-R1 is a specialized LLM based on QwQ-32B, fine-tuned on 1,232 scientific publications and 33,269 candidate materials to discover precursor additives optimizing perovskite solar cells. The model generates solutions for defect passivation and improves stability/performance, experimentally validated.
Read source
Your take?
QwenFine-tuningReasoning

Summary generated by Claude — human-verified