The Finance Lab Introduces TFL Bloodhound, a Financial Reasoning Model Trained on Market Outcomes Instead of Human PreferenceTFL_Bloodhound replaces the human evaluator in RLHF with realized market outcomes, and separates quantitative estimation from language reasoning in a compound architecture. In pilot deployment across energy futures; a public playground is open for testing.
