mukaj/Llama-3.1-Hawkish-8B

mukaj/Llama-3.1-Hawkish-8B is an 8 billion parameter Llama 3.1-based language model fine-tuned for financial topics, economics, and mathematics, with a 32K context length. It was trained on 50 million high-quality financial tokens, alongside general knowledge, coding, and NLP data to mitigate catastrophic forgetting. This model demonstrates significant improvements in financial knowledge and reasoning, notably achieving a passing score on the CFA Level 1 exam, outperforming larger models in this domain.