いつサンプリングするかを学ぶ: 効率的なLLMコト推論のための信頼度認識自己一貫性

arXiv cs.CL / 2026/3/11

Ideas & Deep AnalysisModels & Research

要点

  • 大規模言語モデル(LLM)はチェーン・オブ・ソート(CoT)推論を用いて高い性能を発揮するが、不必要に長い推論経路を生成し、推論コストが増加することが多い。
  • 自己一貫性(self-consistency)手法は複数の推論経路を集約して精度を向上させるが、大幅な計算オーバーヘッドを伴う。
  • 本論文では、単一の推論軌跡からの不確実性推定に基づいて、単一路線推論と多重路線推論を適応的に選択する信頼度認識型意思決定フレームワークを提案する。
  • このフレームワークはMedQAデータで訓練され、追加のファインチューニングなしにMathQA、MedMCQA、MMLUなど他のデータセットでも良好に一般化し、多重路線法と同等の精度を保ちつつトークン使用量を最大80%削減する。
  • この手法は推論経路からの中間シグナルを活用し、LLMの推論精度と推論コストのバランスを効率的かつ転移可能な形で実現する。

Computer Science > Computation and Language

arXiv:2603.08999 (cs)
[Submitted on 9 Mar 2026]

Title:Learning When to Sample: Confidence-Aware Self-Consistency for Efficient LLM Chain-of-Thought Reasoning

View a PDF of the paper titled Learning When to Sample: Confidence-Aware Self-Consistency for Efficient LLM Chain-of-Thought Reasoning, by Juming Xiong and 8 other authors
View PDF HTML (experimental)
Abstract:Large language models (LLMs) achieve strong reasoning performance through chain-of-thought (CoT) reasoning, yet often generate unnecessarily long reasoning paths that incur high inference cost. Recent self-consistency-based approaches further improve accuracy but require sampling and aggregating multiple reasoning trajectories, leading to substantial additional computational overhead. This paper introduces a confidence-aware decision framework that analyzes a single completed reasoning trajectory to adaptively select between single-path and multi-path reasoning. The framework is trained using sentence-level numeric and linguistic features extracted from intermediate reasoning states in the MedQA dataset and generalizes effectively to MathQA, MedMCQA, and MMLU without additional fine-tuning. Experimental results show that the proposed method maintains accuracy comparable to multi-path baselines while using up to 80\% fewer tokens. These findings demonstrate that reasoning trajectories contain rich signals for uncertainty estimation, enabling a simple, transferable mechanism to balance accuracy and efficiency in LLM reasoning.
Subjects: Computation and Language (cs.CL)
Cite as: arXiv:2603.08999 [cs.CL]
  (or arXiv:2603.08999v1 [cs.CL] for this version)
  https://doi.org/10.48550/arXiv.2603.08999
Focus to learn more
arXiv-issued DOI via DataCite

Submission history

From: Juming Xiong [view email]
[v1] Mon, 9 Mar 2026 22:34:06 UTC (273 KB)
Full-text links:

Access Paper:

    View a PDF of the paper titled Learning When to Sample: Confidence-Aware Self-Consistency for Efficient LLM Chain-of-Thought Reasoning, by Juming Xiong and 8 other authors
  • View PDF
  • HTML (experimental)
  • TeX Source
Current browse context:
cs.CL
< prev   |   next >
Change to browse by:
cs

References & Citations

export BibTeX citation Loading...

BibTeX formatted citation

×
Data provided by:

Bookmark

BibSonomy logo Reddit logo
Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
Links to Code Toggle
Papers with Code (What is Papers with Code?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos

Demos

Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers

Recommenders and Search Tools

Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.