Fact-Level Relevance-Aware Confidence Calibration for Large Language Models

The code is implemented based on "LongFact"

python -m main.pipeline

python -m eval.run_eval --result_path=

python -m main.confidence --result_path=