Search papers, labs, and topics across Lattice.
The paper introduces AGSC, a novel Uncertainty Quantification (UQ) framework for long-form generation that addresses the challenges of hallucination and reliable aggregation in LLMs. AGSC uses NLI neutral probabilities to filter irrelevant information and applies Gaussian Mixture Model (GMM) soft clustering to model latent semantic themes for topic-aware weighting. Experiments on BIO and LongFact datasets demonstrate that AGSC achieves state-of-the-art correlation with factuality while reducing inference time by approximately 60% compared to full atomic decomposition methods.
LLMs can be made more reliable and efficient by adaptively focusing uncertainty quantification on relevant semantic themes, cutting inference time by 60% while improving factuality correlation.
Large Language Models (LLMs) have demonstrated impressive capabilities in long-form generation, yet their application is hindered by the hallucination problem. While Uncertainty Quantification (UQ) is essential for assessing reliability, the complex structure makes reliable aggregation across heterogeneous themes difficult, in addition, existing methods often overlook the nuance of neutral information and suffer from the high computational cost of fine-grained decomposition. To address these challenges, we propose AGSC (Adaptive Granularity and GMM-based Semantic Clustering), a UQ framework tailored for long-form generation. AGSC first uses NLI neutral probabilities as triggers to distinguish irrelevance from uncertainty, reducing unnecessary computation. It then applies Gaussian Mixture Model (GMM) soft clustering to model latent semantic themes and assign topic-aware weights for downstream aggregation. Experiments on BIO and LongFact show that AGSC achieves state-of-the-art correlation with factuality while reducing inference time by about 60% compared to full atomic decomposition.
