AI Paper Atlas

Research map / Probabilistic and scientific modeling / Variational and generative sampling

Cache Coherent Resampling for Efficient Test Time Scaling in LLM Reasoning via Adaptive Sequential Monte Carlo

KE WANG, ZEHAO YU, Luwei Wang, Yongchao Huang

ICML 2026 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Sequential Monte Carlo for inference

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace sequential MCMC with ASMC at 128x budgetASMC · MATH500p95 latency (s): 1317.5 -> 75.4Improved
replace sequential MCMC with ASMC at 128x budgetASMC · MATH500accuracy (%): 73.8 -> 80.2Improved
use KV reorder instead of full cache rebuild for resampling (ablation)cache-coherent ASMC · single resampling event, Qwen2.5-Math-7Bevent latency: 53 to 77x fasterImproved
use KV reorder instead of rebuild at L=3072, R=50 (ablation)cache-coherent ASMC · feasibility, N>=80feasibility: rebuild OOM at N>=80; reorder feasible through N=96Improved
use adaptive-N instead of fixed-N ASMC at 16x budget (ablation)ASMC · MATH500accuracy (%): 77.6 -> 76.4Hurt
use adaptive-N instead of fixed-N ASMC at 128x budget (ablation)ASMC · MATH500accuracy (%): 80.2 -> 80.6Improved
switch from multinomial to residual resampling (ablation)ASMC · MATH500 (200 problems)accuracy (%): +0.5%No clear effect
vary ESS threshold from 0.3 to 0.7 (ablation)ASMC-adaptive · MATH500accuracy (%): 79.4 -> 79.0No clear effect
more resampling events (>=16 vs 1-2) (ablation)ASMC · MATH500accuracy (%): 100.0% -> 43.9%Hurt
increase best-of-n from n=4 to larger n (log-prob reranking) (ablation)Best-of-n · MATH500accuracy (%): 79.2 at n=4No clear effect
use ASMC (N=4) instead of greedy at 8x budget on codeASMC · HumanEvalpass@1 (%): 55.5 -> 49.4Hurt
scale ASMC to N=64 at 128x budget on codeASMC · HumanEvalpass@1 (%): 56.1 -> 64.0Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy