BIS paper explores language models for central banking (CB-LMs).
By Bank for International Settlements (BIS)
This working paper introduces CB-LMs, specialised encoder-only models trained on a broad range of central bank texts. CB-LMs outperform their foundational models in predicting central bank idioms and even exceed state-of-the-art generative large language models (LLMs) in classifying monetary policy stances from Federal Open Market Committee (FOMC) statements.
- Use cases, geography and tags
- Organizations that created, adopted or are mentioned
- Ecosystem position
- Link to the original asset
- Comments and reactions