Skip to main content
Search

# Power Sampling

3 related articles

Subscribe to updates
  1. arXiv

    Power sampling no longer takes one step per call: prefetching the accept/reject subtree cuts PowerMH time by a median 1.71×

    PowerMH must decide each accept/reject before it can generate the next proposal; PreSTO prefetches several future proposals in one batched call, giving a median 1.71× speedup over PowerMH across seven datasets and five base LLMs while leaving its transition kernel unchanged.

    Morgan Stanley (United States)Georgia Institute of TechnologyThe University of Texas at El Paso
    Nan Jiang（First author）
    PreSTOPower SamplingMarkov Chain Monte CarloInference accelerationPredictive Traversal Rule
  2. arXiv

    Power sampling fails to converge at real compute budgets, and simple Best-of-N matches its reasoning gains

    MCMC and SMC samplers for the power distribution fail to converge under realistic compute budgets, and their reasoning gains are largely recreated by simple Best-of-N selection; Power-RLOO, which fine-tunes directly toward the power distribution, helps only when the base model is already aligned with the task.

    Lancaster University
    Abiel Malkani Talwar（First author）
    Power-RLOOPower SamplingSequential Monte CarloMarkov Chain Monte CarloTraining-Free LLM Reasoning
  3. arXiv

    Power-SMC targets the sequence-level power distribution with parallel candidates and importance-weight pruning, matching or exceeding MH sampling on MATH500, GSM8K, GPQA and HumanEval while accelerating inference by up to 17.6x

    The work introduces Power-SMC, a training-free sampling method that maintains multiple candidate sequences in parallel, scores each with an importance weight measuring how well it matches the sequence-level power distribution, and periodically prunes low-scoring candidates in favor of high-scoring ones, matching or exceeding Metropolis-Hastings sampling in accuracy on MATH500, GSM8K, GPQA and HumanEval, preserving output diversity, and accelerating inference by up to 17.6x.

    Intel (United States)University of Southern California
    Seyedarmin Azizi（First author）
    Power-SMCPower SamplingTraining-Free LLM Reasoning

You’re all caught up