Argonne researchers develop framework to make LLMs safer for scientific applications
Most existing benchmarks were designed for general-purpose AI applications, however, and often overlook the unique challenges of scientific research
Credit: Andrey Suslov/Shutterstock
New approach combines domain knowledge, retrieval and model safeguards to reduce the risk of unreliable AI-generated outputs in scientific computing workflows
Register for FREE to keep reading
Join 12,000+ scientists, engineers, and IT professionals driving innovation through informatics, HPC, and simulation with:
- Insights into HPC, AI, lab informatics & data
- Curated content for life sciences, engineering & academia
- Access to Breakthroughs: real-world computing success
- Free reports & panels, including the Lab Informatics Guide
- White Papers & software updates for smarter research
Sign up now
Already a member? Log in here
Your data is protected under our privacy policy.
