We built Big Finance Bench to help define what good looks like for AI in finance. Now, we’re extending that evaluation framework to financial data.
We’re focused on two questions: how can LLMs use individual financial data sources more effectively, and how should multiple sources work together to surface the right data for each task?
Early results from these optimizations show 75% fewer serious hallucinations, 40% lower latency, and a 4x reduction in token consumption.
We’re focused on two questions: how can LLMs use individual financial data sources more effectively, and how should multiple sources work together to surface the right data for each task?
Early results from these optimizations show 75% fewer serious hallucinations, 40% lower latency, and a 4x reduction in token consumption.
2 6 0 29 1.4K 5