uncertainty calibration

2 items

RESEARCHarXiv CS.CL·4/15/2026

Think Through Uncertainty: Improving Long-Form Generation Factuality via Reasoning Calibration

This research introduces CURE, a novel framework designed to improve the factuality of long-form generation by LLMs by teaching them to reason about uncertainty at the claim level. It aims to overcome the limitation of models often stating incorrect claims confidently, focusing instead on granular uncertainty calibration.

LLMs hallucination uncertainty calibration Reasoning

RESEARCHarXiv CS.LG·20d ago

UCCI: Calibrated Uncertainty for Cost-Optimal LLM Cascade Routing

UCCI is an innovative router that uses calibrated uncertainty to optimize the cost of LLM cascades, sending easy queries to smaller models and difficult ones to larger models. It reduces inference cost by 31% on production workloads while maintaining accuracy, by calibrating model confidence.

LLM routing uncertainty calibration model cascades Cost Optimization