Tag
This paper presents a language model forecasting system for merger arbitrage that combines expert-guided context engineering with fine-tuning on historical deals, achieving state-of-the-art performance on over 400 large deals across 42 countries.