Lune

ACL Findings2026

Adaptive Test-Time Compute Allocation with Evolving In-Context Demonstrations

Bowen Zuo, Dongruo Zhou, Yinglun Zhu

Abstract

While scaling test-time compute can substantially improve model performance, existing approaches either rely on static compute allocation or sample from fixed generation distributions. In this work, we introduce a testtime compute allocation framework that jointly adapts where computation is spent and how generation is performed. Our method begins with a warm-up phase that identifies easy queries and assembles an initial pool of question-response pairs from the test set itself. An adaptive phase then concentrates further computation on unresolved queries while reshaping their generation distributions through evolving in-context demonstrationsconditioning each generation on successful responses from semantically related queries rather than resampling from a fixed distribution. Experiments across math, coding, and reasoning benchmarks demonstrate that our approach consistently outperforms existing baselines while consuming substantially less inference-time compute.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

the full paper · related work

Get the answer

Free to start. No credit card required.

Lineage

  • Tree of Thoughts: Deliberate Problem Solving with Large Language Models cites Self-Refine: Iterative Refinement with Self-Feedback
  • Tree of Thoughts: Deliberate Problem Solving with Large Language Models cites Language Models are Few-Shot Learners
  • Tree of Thoughts: Deliberate Problem Solving with Large Language Models cites Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
  • Self-Refine: Iterative Refinement with Self-Feedback cites Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
  • Large Language Models are Zero-Shot Reasoners cites Language Models are Few-Shot Learners
  • Large Language Models are Zero-Shot Reasoners cites Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
Builds onCited byLeft to right by year; the more cited, the higher.

Builds on21

  1. Language Models are Few-Shot LearnersTom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah et al.NeurIPS 2020 · 64,255 citations
  2. Chain-of-Thought Prompting Elicits Reasoning in Large Language ModelsJason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma et al.NeurIPS 2022 · 22,562 citations
  3. Large Language Models are Zero-Shot ReasonersTakeshi Kojima, Shixiang Shane Gu, Machel Reid, Yutaka Matsuo et al.NeurIPS 2022 · 8,168 citations
  4. Tree of Thoughts: Deliberate Problem Solving with Large Language ModelsShunyu Yao, Dian Yu, Jeffrey Zhao, Izhak Shafran et al.NeurIPS 2023 · 5,068 citations
  5. Self-Refine: Iterative Refinement with Self-FeedbackAman Madaan, Niket Tandon, Prakhar Gupta, Skyler Hallinan et al.NeurIPS 2023 · 4,972 citations

Related papers

  1. Strategic Scaling of Test-Time Compute: A Bandit Learning ApproachBowen Zuo, Yinglun ZhuICLR 2026 · 9 citations
  2. What If We Allocate Test-Time Compute Adaptively?Ahsan Bilal, Muhammad Ahmed Mohsin, Muhammad Umer, Ali Subhan et al.ICML 2026 · 3 citations
  3. A Reward-Guided Dual-Phase Framework for Adaptive Inference-Time ReasoningYingqian Cui, Zhenwei Dai, Pengfei He, Bing He et al.ACL Findings 2026
  4. RTTC: Reward-Guided Collaborative Test-Time ComputeJuan Pablo Muñoz, Jinjie YuanEMNLP Findings 2025
  5. SCALE: Selective Resource Allocation for Overcoming Performance Bottlenecks in Mathematical Test-time ScalingYang Xiao, Chunpu Xu, Ruifeng Yuan, Jessie Wang et al.AAAI 2026 · 1 citation
Dusk over the sea between two cliffs drawn in fine vertical lines