Inference-Time Compute Allocation

Follow this topic Ask about this topic

Momentum

10 papers in the last four weeks, against 2 the four weeks before. 0.1% of all new papers.

Jul 13Week of Sep 28

Latest papers 53

All topics

Readout Stability in Prefill-Only Decision Models:Zero-Label Prediction and Inference-Time Compute Allocation
Oct 6, 2026Ran Li, Lei ChenInference-Time Compute Allocation Efficient Language Model Inference
Hong Kong University of Science and Technology · Hong Kong University of Science and Technology(Guangzhou)
Expanding LLM Reasoning
Oct 4, 2026Rian Atri, Evan LuoLLM Inference Efficiency Inference-Time Scaling
Keiji AI · University of California, Berkeley
ReSolve: Reusing Candidate Reasoning through Selective Generative Moderation
Oct 1, 2026Bangji Yang, Jiajun Fan, Hongba Ma +6Efficient Language Model Reasoning Inference-Time Compute Allocation
University of Illinois at Urbana-Champaign · Tsinghua University · Rutgers University +2
Budget Boundary Effects in Test-Time Mathematical Reasoning
Sep 30, 2026Guilin Zhang, Ziqi Tan, Wulan Guo +5Mathematical Reasoning Benchmarks Test-Time Optimization
Yantong AI
Thinking Before Thinking: Scaling Agentic Inference Through Meta-Reasoning
Sep 29, 2026Paras Dahal, Anton Bakhtin, Taco Cohen +9Agentic Reasoning Inference-Time Compute Allocation
Meta Superintelligence Labs
The Decision Value of Perception Compute
Sep 28, 2026Hoang Pham Cong, Ho Viet Duc LuongCost-Aware Inference Autonomous Driving Perception
Hanoi University of Science and Technology (HUST), Hanoi, Vietnam
ReSync: Re-Aligning the Two Clocks of Asynchronous World-Action Models
Sep 27, 2026Xi Lin, Feihong Zhang, Yulong Shi +6Action-Conditioned Video Generation Video Diffusion Models
Johns Hopkins University · Tsinghua University · Yinwang Intelligent Technology Co., Ltd. +1
The Organization of Inference: Information, Resource Constraints, and AI Production
Sep 17, 2026Yukun Zhang, Kemu Xu, Yishen ChenLLM Planning Inference-Time Scaling
The Chinese University of Hong Kong Hong Kong, China · University of Edinburgh Edinburgh, United Kingdom · The Chinese University of Hong Kong, Shenzhen Shenzhen, China
When Agents Slow Down: Understanding LLM Agents' Test-Time Strategies via Elo-per-token Analysis
Sep 14, 2026Kaiyuan Liu, Qiuyang Mang, Bo Peng +6LLM Agent Evaluation Test-Time Scaling
UC Berkeley · University of Washington · Princeton University +1
Intelligence Under Time Constraints: Rethinking Test-Time Compute
Sep 14, 2026Xiaotian ZhangInference-Time Scaling Inference-Time Compute Allocation
Trooly.AI
Routing by Reasoning Need: Trajectory-Aware Decoding Control for Diffusion Vision-Language Models
Sep 12, 2026Yixiang Liu, Zhongxing Xu, Zhonghua Wang +1Diffusion Language Model Inference VLM Reasoning
Southern University of Science and Technology · Monash University
Keep, Customize, or Exit: Default Design and Token Pricing in LLM Reasoning Services
Aug 13, 2026Ahmet Bugra Gundogan, Yigit Turkmen, Melih BastopcuLLM Serving Inference-Time Compute Allocation
Department of Electrical and Electronics Engineering
Sampling Luck Masquerades as Allocation Gain: Auditing Test-Time Budget Allocation for Neural Combinatorial Optimization
Aug 13, 2026Jinhyung BaeAlgorithmic Auditing Inference-Time Compute Allocation
Department of Industrial and Management Engineering and the AI Data Convergence program, Hankuk University of Foreign Studies, Yongin, Republic of Korea
Thought-Level Beam Search for Reasoning
Aug 8, 2026Lijie Yang, Hongyin Luo, Jiawei Zhao +2Inference-Time Search Efficient Language Model Reasoning
Princeton University · MIT CSAIL · Meta AI
Thinking Hard, Not Smart: Reasoning Models Fail to Ration Test-Time Compute Across Questions
Aug 8, 2026Chenrui Fan, Yize Cheng, Ming Li +3LLM Evaluation Inference-Time Compute Allocation
University of Maryland, College Park · MBZUAI, UAE
CoBa: Cost-Effective Test-Time Scaling via Compute-Balanced Routing
Aug 7, 2026Yan Zhou, Yue Ouyang, Kaiyang Zheng +1Cost-Aware Inference Test-Time Scaling
School of Mathematics and Statistics, Changsha University of Science and Technology, Changsha, China · School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai, China.
Rethinking Inference-Time Scaling in Local Computer-Use Agents: Failure Modes and Compute Tradeoffs
Jul 30, 2026Woongkyu Lee, Jungwook ChoiComputer-Use Agents Inference-Time Scaling
Hanyang University
Search as Computation Allocation
Jul 30, 2026Alexander TuisovInference-Time Compute Allocation
Technion - Israel Institute of Technology
Exploring Budgeted Image Classification with Content-Sensitive Resource Allocation
Jul 27, 2026Athanasios G. PapadopoulosCost-Aware Inference Classification
Tandon School of Engineering, New York University
Resample or Reroute? Recoverable Stopping Debt Without Identified Action Selection
Jul 9, 2026Teng-Ruei ChenCode Generation LLM Reliability
Institute of Bioinformatics and Systems Biology, National Yang Ming Chiao Tung University, Hsinchu, Taiwan · Krixvon, Taipei, Taiwan
ELASTIC: Efficiently Learning to Adaptively Scale Test-Time Compute for Generative Control Policies
Jun 30, 2026Andrew Zou Li, Gokul Swamy, Yonatan Bisk +1Diffusion Policy Test-Time RL
Carnegie Mellon University
End-to-End Dynamic Sparsity for Resource-Adaptive LLM Inference
Jun 26, 2026Yuhang Chen, Jinhao Duan, Ruichen Zhang +11LLM Inference Structured Sparsity
University of North Carolina at Chapel Hill · Meta AI
Humans Disengage, Reasoning Models Persist: Separating Difficulty Registration from Deliberation Allocation
Jun 25, 2026Han-yu WangComputational Cognitive Modeling Inference-Time Compute Allocation
The University of Hong Kong
Maestro Order: A Model-Agnostic Orchestration Harness
Jun 22, 2026Hidayet AksuLLM Agent Orchestration LLM Reliability
Heteroskedastic Signals in Budgeted LLM Verification: Structural Heterogeneity Limits Optimization Gains
Jun 14, 2026Jinlong YangLarge Language Model-Guided Optimization LLM Uncertainty Estimation
Northwestern Polytechnical University
DIRECT: When and Where Should You Allocate Test-Time Compute in Embodied Planners?
Jun 10, 2026Jadelynn Dao, Milan Ganai, Yasmina Abukhadra +7Test-Time Scaling for VLMs Inference-Time Scaling
Stanford University · University of Waterloo · NVIDIA
Intrinsic Selection and Particle Resampling for Inference-Time Scaling Beyond Domain Verifiability
Jun 7, 2026Giorgio Giannone, Mustafa Eyceoz, Shabana Baig +5Particle Filtering Inference-Time Scaling
AI Innovation, Red Hat · DeCoDE Lab, MIT · Core AI, IBM
When to Think Deeply: Inhibitory Deliberation for LLM Reasoning
Jun 4, 2026Zhixuan He, Yue FengEfficient Language Model Reasoning Inference-Time Compute Allocation
University of Birmingham, United Kingdom
Not All Errors Are Equal: Consequence-Aware Reasoning Compute Allocation
Jun 3, 2026Liang He, Jingbo Wen, Haoyu Wang +4Cost-Aware Inference Inference-Time Compute Allocation
Shanghai Institute of Optics and Fine Mechanics · The University of Sydney
The Shadow Price of Reasoning: Economic Perspective on Optimal Budget Allocation for LLMs
Jun 2, 2026Xu Wan, Speed Zhu, Jianwei Cai +4Cost-Aware Inference Inference-Time Scaling
1Zhejiang University · 2Tecent HY Team · 3Peking University

Oct 6, 2026, cs.CLJ/K move · Enter open · S save

Readout Stability in Prefill-Only Decision Models:Zero-Label Prediction and Inference-Time Compute Allocation

Ran Li, Lei Chen

Hong Kong University of Science and Technology · Hong Kong University of Science and Technology(Guangzhou)

Inference-Time Compute Allocation Efficient Language Model Inference

Open paper AskPDF arXiv

Prefill-only decision models inspired by the Jev model score every candidate in a menu during a single forward pass and never decode, which makes one call one to two orders of magnitude cheaper than a same-scale generative language model. We show that this read-out structure comes with a testable property. When an intervention changes only the candidate menu and leaves the input text fixed, the post-intervention accuracy is already determined by the cached first-pass distribution. The estimator restricts the pass-1 probabilities to the menu, renormalizes, and reads off the argmax; it uses no labels and no second forward pass. Across seven model families, ten datasets and two task types, menu-only interventions are predicted to within 4.2 points, and for one family the prediction is exact. A probability-level variant of the same estimator errs by 21.0 points, so the property lives in the ranking rather than in the probabilities and is not recovered by calibration. Same-scale generative language models do not share the property. On those models the same estimator errs by 1.6 to 15.8 points and degrades as the model grows. The property turns inference-time compute into a decision that can be made before deployment. Uniform extra passes buy calibration but almost no accuracy; at matched cost a confidence cascade outperforms every scheme that re-asks the same model, and curating the menu beats enlarging the model, with a 0.8B model on a curated 5-candidate menu reaching 95.4% on CLINC150 against 80.0% for a 4B model on the full 150-label menu.Code and data are available at https://github.com/rlisml/jev-cascade.

Momentum

Related topics

Latest papers 53