Saved in:
Bibliographic Details
Main Authors: Lin, Kevin, Snell, Charlie, Wang, Yu, Packer, Charles, Wooders, Sarah, Stoica, Ion, Gonzalez, Joseph E.
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2504.13171
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910913574993920
author Lin, Kevin
Snell, Charlie
Wang, Yu
Packer, Charles
Wooders, Sarah
Stoica, Ion
Gonzalez, Joseph E.
author_facet Lin, Kevin
Snell, Charlie
Wang, Yu
Packer, Charles
Wooders, Sarah
Stoica, Ion
Gonzalez, Joseph E.
contents Scaling test-time compute has emerged as a key ingredient for enabling large language models (LLMs) to solve difficult problems, but comes with high latency and inference cost. We introduce sleep-time compute, which allows models to "think" offline about contexts before queries are presented: by anticipating what queries users might ask and pre-computing useful quantities, we can significantly reduce the compute requirements at test-time. To demonstrate the efficacy of our method, we create modified versions of two reasoning tasks - Stateful GSM-Symbolic and Stateful AIME. We find that sleep-time compute can reduce the amount of test-time compute needed to achieve the same accuracy by ~ 5x on Stateful GSM-Symbolic and Stateful AIME and that by scaling sleep-time compute we can further increase accuracy by up to 13% on Stateful GSM-Symbolic and 18% on Stateful AIME. Furthermore, we introduce Multi-Query GSM-Symbolic, which extends GSM-Symbolic by including multiple related queries per context. By amortizing sleep-time compute across related queries about the same context using Multi-Query GSM-Symbolic, we can decrease the average cost per query by 2.5x. We then conduct additional analysis to understand when sleep-time compute is most effective, finding the predictability of the user query to be well correlated with the efficacy of sleep-time compute. Finally, we conduct a case-study of applying sleep-time compute to a realistic agentic SWE task.
format Preprint
id arxiv_https___arxiv_org_abs_2504_13171
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Sleep-time Compute: Beyond Inference Scaling at Test-time
Lin, Kevin
Snell, Charlie
Wang, Yu
Packer, Charles
Wooders, Sarah
Stoica, Ion
Gonzalez, Joseph E.
Artificial Intelligence
Computation and Language
Scaling test-time compute has emerged as a key ingredient for enabling large language models (LLMs) to solve difficult problems, but comes with high latency and inference cost. We introduce sleep-time compute, which allows models to "think" offline about contexts before queries are presented: by anticipating what queries users might ask and pre-computing useful quantities, we can significantly reduce the compute requirements at test-time. To demonstrate the efficacy of our method, we create modified versions of two reasoning tasks - Stateful GSM-Symbolic and Stateful AIME. We find that sleep-time compute can reduce the amount of test-time compute needed to achieve the same accuracy by ~ 5x on Stateful GSM-Symbolic and Stateful AIME and that by scaling sleep-time compute we can further increase accuracy by up to 13% on Stateful GSM-Symbolic and 18% on Stateful AIME. Furthermore, we introduce Multi-Query GSM-Symbolic, which extends GSM-Symbolic by including multiple related queries per context. By amortizing sleep-time compute across related queries about the same context using Multi-Query GSM-Symbolic, we can decrease the average cost per query by 2.5x. We then conduct additional analysis to understand when sleep-time compute is most effective, finding the predictability of the user query to be well correlated with the efficacy of sleep-time compute. Finally, we conduct a case-study of applying sleep-time compute to a realistic agentic SWE task.
title Sleep-time Compute: Beyond Inference Scaling at Test-time
topic Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2504.13171