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

• URL: http://arxiv.org/abs/2605.15100v1
• Date: Thu, 14 May 2026 17:19:37 GMT
• Title: Dual-Dimensional Consistency: Balancing Budget and Quality in Adaptive Inference-Time Scaling
• Abstract summary: Large Language Models (LLMs) have demonstrated remarkable abilities in reasoning.<n> maximizing their potential through inference-time scaling faces challenges in trade-off between sampling budget and reasoning quality.<n>We propose Dual-Dimensional Consistency (DDC), a unified framework that bridges path quality with adaptive termination.
• Score: 12.49034574686497
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• 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.

Related papers

• SFAD: Speculative Factuality-Aware Decoding [13.350717038599386]
  ConFide is a preference dataset with fine-grained atomic perturbations.<n>Epistemic Friction detects potential hallucinations by quantifying distributional tension weighted by specialist certainty.<n>Asymmetric Logit Steering refines the target distribution through residual-based logit injection.
  arXiv  Detail & Related papers  (2026-09-01T06:46:54Z)
• Is Your Model Thinking or Just Stagnating? PUMA: Diagnosing Reasoning Pathology via Phase-Momentum Alignment [13.9489759632504]
  Test-time scaling empowers Large Reasoning Models (LRMs) to tackle complex tasks via extensive Chain-of-Thought (CoT)<n>This paper introduces PUMA, a training-free framework employing a tiered diagnostic architecture.<n>Experiments on LRMs spanning 1.5B to 32B demonstrate that PUMA consistently outperforms state-of-the-art baselines across diverse benchmarks.
  arXiv  Detail & Related papers  (2026-07-19T11:00:59Z)
• Think Less, Know More: State-Aware Reasoning Compression with Knowledge Guidance for Efficient Reasoning [16.46227355517168]
  Large Reasoning Models (LRMs) achieve strong performance on complex tasks by leveraging long Chain-of-Thought (CoT)<n>Existing CoT compression methods struggle to balance accuracy and efficiency, and lack fine-grained, step-level adaptation to redundancy and reasoning bias.<n>We propose State-Aware Reasoning Compression with Knowledge Guidance ( Experiments), a framework that performs step-wise CoT compression.
  arXiv  Detail & Related papers  (2026-04-10T09:31:41Z)
• ECHO: Elastic Speculative Decoding with Sparse Gating for High-Concurrency Scenarios [12.103618967955846]
  Speculative Decoding promises to accelerate the inference of Large Language Models, yet its efficacy often degrades in production-grade serving.<n>We introduce ECHO, a high-oriented framework integrated into SGLang that reformulates speculative execution as a budgeted scheduling problem.<n>ECHO consistently outperforms SOTA methods in both low-load and high-load scenarios, achieving up to 5.35x walltime speedup and delivering over 20% relative speedup gain.
  arXiv  Detail & Related papers  (2026-03-10T03:51:24Z)
• Hit-RAG: Learning to Reason with Long Contexts via Preference Alignment [97.55382322103986]
  Hit-RAG is a multi-stage preference alignment framework designed to resolve cognitive bottlenecks.<n>Our approach systematically refines the utilization of external evidence via three distinct stages.
  arXiv  Detail & Related papers  (2026-03-07T04:05:27Z)
• Stable Adaptive Thinking via Advantage Shaping and Length-Aware Gradient Regulation [14.501114943020589]
  Large reasoning models (LRMs) achieve strong performance through extended reasoning traces.<n>LRMs often exhibit overthinking behavior for low-complexity queries.<n>We propose a two-stage framework for stable adaptive thinking in LRMs.
  arXiv  Detail & Related papers  (2026-02-26T02:49:36Z)
• Beyond Confidence: Adaptive and Coherent Decoding for Diffusion Language Models [64.92045568376705]
  Coherent Contextual Decoding (CCD) is a novel inference framework built upon two core innovations.<n>CCD employs a trajectory rectification mechanism that leverages historical context to enhance sequence coherence.<n>Instead of rigid allocations based on diffusion steps, we introduce an adaptive sampling strategy that dynamically adjusts the unmasking budget for each step.
  arXiv  Detail & Related papers  (2025-11-26T09:49:48Z)
• Efficient Thought Space Exploration through Strategic Intervention [54.35208611253168]
  We propose a novel Hint-Practice Reasoning (HPR) framework that operationalizes this insight through two synergistic components.<n>The framework's core innovation lies in Distributional Inconsistency Reduction (DIR), which dynamically identifies intervention points.<n> Experiments across arithmetic and commonsense reasoning benchmarks demonstrate HPR's state-of-the-art efficiency-accuracy tradeoffs.
  arXiv  Detail & Related papers  (2025-11-13T07:26:01Z)
• Perception-Consistency Multimodal Large Language Models Reasoning via Caption-Regularized Policy Optimization [72.30168853571216]
  multimodal large language models excel at tasks that integrate visual perception with symbolic reasoning.<n>CapPO integrates two key mechanisms: (1) a caption-based consistency regularization, which minimizes the divergence between responses conditioned on raw images and those conditioned on captions, and (2) a KL-weighted advantage estimation scheme, which adaptively scales reinforcement signals to strengthen perceptually consistent trajectories.
  arXiv  Detail & Related papers  (2025-09-26T04:32:26Z)
• Fractured Chain-of-Thought Reasoning [61.647243580650446]
  We introduce Fractured Sampling, a unified inference-time strategy that interpolates between full CoT and solution-only sampling.<n>We show that Fractured Sampling consistently achieves superior accuracy-cost trade-offs, yielding steep log-linear scaling gains in Pass@k versus token budget.
  arXiv  Detail & Related papers  (2025-05-19T11:30:41Z)
• ConCISE: Confidence-guided Compression in Step-by-step Efficient Reasoning [64.93140713419561]
  Large Reasoning Models (LRMs) perform strongly in complex reasoning tasks via Chain-of-Thought (CoT) prompting, but often suffer from verbose outputs.<n>Existing fine-tuning-based compression methods either operate post-hoc pruning, risking disruption to reasoning coherence, or rely on sampling-based selection.<n>We introduce ConCISE, a framework designed to generate concise reasoning chains, integrating Confidence Injection to boost reasoning confidence, and Early Stopping to terminate reasoning when confidence is sufficient.
  arXiv  Detail & Related papers  (2025-05-08T01:40:40Z)
• False Correlation Reduction for Offline Reinforcement Learning [115.11954432080749]
  We propose falSe COrrelation REduction (SCORE) for offline RL, a practically effective and theoretically provable algorithm. We empirically show that SCORE achieves the SoTA performance with 3.1x acceleration on various tasks in a standard benchmark (D4RL)
  arXiv  Detail & Related papers  (2021-10-24T15:34:03Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.