Inducing Group Fairness in LLM-Based Decisions