Dual-Dimensional Consistency: Balancing Budget and Quality in Adaptive Inference-Time Scaling

TL;DR

This paper introduces Dual-Dimensional Consistency, a framework for adaptive inference in LLMs that balances reasoning quality and resource efficiency, reducing token use significantly while maintaining accuracy.

Contribution

It presents a novel unified approach combining confidence-weighted Bayesian and trend-aware pruning to improve inference efficiency and reasoning quality in LLMs.

Findings

Reduces token consumption by over 10 times.

Maintains or exceeds baseline accuracy across five benchmarks.

Effectively filters hallucinations while accelerating reasoning.

Abstract

Large Language Models (LLMs) have demonstrated remarkable abilities in reasoning. However, maximizing their potential through inference-time scaling faces challenges in trade-off between sampling budget and reasoning quality. Current strategies remain inefficient as they typically treat sampling width and depth as orthogonal objectives, where width consensus methods risk reinforcing hallucinations, while depth pruning mechanisms prematurely truncate complex yet valid reasoning chains. Therefore, we propose Dual-Dimensional Consistency (DDC), a unified framework that bridges path quality with adaptive termination. By coupling Confidence-Weighted Bayesian protocol with a Trend-Aware Stratified Pruning, our method ensures that computational resources are concentrated on high quality reasoning paths, filtering hallucinations while accelerating consensus. Evaluations across five benchmarks demonstrate that this approach reduces token consumption by over 10 times while maintaining or exceeding the accuracy of strong baselines across various LLMs.