Calibrating LLM Confidence with Semantic Steering: A Multi-Prompt Aggregation Framework

Open in new window