Cover Image

Beyond Test-Time Compute Strategies: Advocating Energy-per-Token in LLM Inference

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Wilhelm, Patrick, Wittkopp, Thorsten, Kao, Odej
Format:	Preprint
Published:	2026
Subjects:	Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Monitoring Emergent Reward Hacking During Generation via Internal Activations
by: Wilhelm, Patrick, et al.
Published: (2026)

LogRCA: Log-based Root Cause Analysis for Distributed Services
by: Wittkopp, Thorsten, et al.
Published: (2024)

Beyond Microservices: Testing Web-Scale RCA Methods on GPU-Driven LLM Workloads
by: Scheinert, Dominik, et al.
Published: (2026)

Revisiting Gradient Staleness: Evaluating Distance Metrics for Asynchronous Federated Learning Aggregation
by: Wilhelm, Patrick, et al.
Published: (2026)

Quantifying the Energy Consumption and Carbon Emissions of LLM Inference via Simulations
by: Özcan, Miray, et al.
Published: (2025)

Comparative Analysis of Large Language Models for the Machine-Assisted Resolution of User Intentions
by: Flerlage, Justus, et al.
Published: (2025)

Carbon-Aware Quality Adaptation for Energy-Intensive Services
by: Wiesner, Philipp, et al.
Published: (2024)

Noise-aware Client Selection for carbon-efficient Federated Learning via Gradient Norm Thresholding
by: Wilhelm, Patrick, et al.
Published: (2026)

Latency and Token-Aware Test-Time Compute
by: Huang, Jenny Y., et al.
Published: (2025)

Sleep-time Compute: Beyond Inference Scaling at Test-time
by: Lin, Kevin, et al.
Published: (2025)

HAMburger: Accelerating LLM Inference via Token Smashing
by: Liu, Jingyu, et al.
Published: (2025)

Beyond Memorization: Testing LLM Reasoning on Unseen Theory of Computation Tasks
by: Shelat, Shlok, et al.
Published: (2026)

Extending Token Computation for LLM Reasoning
by: Liao, Bingli, et al.
Published: (2024)

DiffAdapt: Difficulty-Adaptive Reasoning for Token-Efficient LLM Inference
by: Liu, Xiang, et al.
Published: (2025)

Empirical Evidences for the Effects of Feature Diversity in Open Set Recognition and Continual Learning
by: Xu, Jiawen, et al.
Published: (2025)

TIDE: Token-Informed Depth Execution for Per-Token Early Exit in LLM Inference
by: Jaber, Jaber, et al.
Published: (2026)

Do Large Language Models Advocate for Inferentialism?
by: Arai, Yuzuki, et al.
Published: (2024)

Reasoning in Token Economies: Budget-Aware Evaluation of LLM Reasoning Strategies
by: Wang, Junlin, et al.
Published: (2024)

SDSAT: Accelerating LLM Inference through Speculative Decoding with Semantic Adaptive Tokens
by: Liu, Chengbo, et al.
Published: (2024)

Can 1B LLM Surpass 405B LLM? Rethinking Compute-Optimal Test-Time Scaling
by: Liu, Runze, et al.
Published: (2025)

Softmax Attention with Constant Cost per Token
by: Heinsen, Franz A.
Published: (2024)

Towards Thinking-Optimal Scaling of Test-Time Compute for LLM Reasoning
by: Yang, Wenkai, et al.
Published: (2025)

Forest-of-Thought: Scaling Test-Time Compute for Enhancing LLM Reasoning
by: Bi, Zhenni, et al.
Published: (2024)

Beyond Tokens: Semantic-Aware Speculative Decoding for Efficient Inference by Probing Internal States
by: Dong, Ximing, et al.
Published: (2026)

A layered architecture for log analysis in complex IT systems
by: Wittkopp, Thorsten
Published: (2025)

Beyond Literal Token Overlap: Token Alignability for Multilinguality
by: Hämmerl, Katharina, et al.
Published: (2025)

SlimInfer: Accelerating Long-Context LLM Inference via Dynamic Token Pruning
by: Long, Lingkun, et al.
Published: (2025)

Mask Tokens as Prophet: Fine-Grained Cache Eviction for Efficient dLLM Inference
by: Huang, Jianuo, et al.
Published: (2025)

LongRoPE: Extending LLM Context Window Beyond 2 Million Tokens
by: Ding, Yiran, et al.
Published: (2024)

Expanding Computation Spaces of LLMs at Inference Time
by: Jang, Yoonna, et al.
Published: (2025)

Informed Routing in LLMs: Smarter Token-Level Computation for Faster Inference
by: Han, Chao, et al.
Published: (2025)

Optimized Multi-Token Joint Decoding with Auxiliary Model for LLM Inference
by: Qin, Zongyue, et al.
Published: (2024)

Rethinking Agentic Workflows: Evaluating Inference-Based Test-Time Scaling Strategies in Text2SQL Tasks
by: Guo, Jiajing, et al.
Published: (2025)

A Survey of Test-Time Compute: From Intuitive Inference to Deliberate Reasoning
by: Ji, Yixin, et al.
Published: (2025)

Moving Beyond Marginal Carbon Intensity: A Poor Metric for Both Carbon Accounting and Grid Flexibility
by: Wiesner, Philipp, et al.
Published: (2025)

TTQ: Activation-Aware Test-Time Quantization to Accelerate LLM Inference On The Fly
by: Koike-Akino, Toshiaki, et al.
Published: (2026)

LazyLLM: Dynamic Token Pruning for Efficient Long Context LLM Inference
by: Fu, Qichen, et al.
Published: (2024)

Compute Optimal Tokenization
by: Limisiewicz, Tomasz, et al.
Published: (2026)

FlexLLM: Token-Level Co-Serving of LLM Inference and Finetuning with SLO Guarantees
by: Oliaro, Gabriele, et al.
Published: (2024)

A Survey on LLM Inference-Time Self-Improvement
by: Dong, Xiangjue, et al.
Published: (2024)