Reward-Shifted Speculative Sampling Is An Efficient Test-Time Weak-to-Strong Aligner

• URL: http://arxiv.org/abs/2508.15044v3
• Date: Tue, 23 Sep 2025 17:25:59 GMT
• Title: Reward-Shifted Speculative Sampling Is An Efficient Test-Time Weak-to-Strong Aligner
• Authors: Bolian Li, Yanran Wu, Xinyu Luo, Ruqi Zhang,
• Abstract summary: We introduce the reward-shifted speculative sampling (SSS) algorithm, in which the draft model is aligned with human preferences, while the target model remains unchanged.<n>Our algorithm achieves superior gold reward scores at a significantly reduced inference cost in test-time weak-to-strong alignment experiments.
• Score: 24.152878302325508
• License: http://creativecommons.org/publicdomain/zero/1.0/
• Abstract: Aligning large language models (LLMs) with human preferences has become a critical step in their development. Recent research has increasingly focused on test-time alignment, where additional compute is allocated during inference to enhance LLM safety and reasoning capabilities. However, these test-time alignment techniques often incur substantial inference costs, limiting their practical application. We are inspired by the speculative sampling acceleration, which leverages a small draft model to efficiently predict future tokens, to address the efficiency bottleneck of test-time alignment. We introduce the reward-shifted speculative sampling (SSS) algorithm, in which the draft model is aligned with human preferences, while the target model remains unchanged. We theoretically demonstrate that the distributional shift between the aligned draft model and the unaligned target model can be exploited to recover the RLHF optimal solution without actually obtaining it, by modifying the acceptance criterion and bonus token distribution. Our algorithm achieves superior gold reward scores at a significantly reduced inference cost in test-time weak-to-strong alignment experiments, thereby validating both its effectiveness and efficiency.

Related papers

• $\ abla$-Reasoner: LLM Reasoning via Test-Time Gradient Descent in Latent Space [71.23672814629448]
  $nabla$-Reasoner is an iterative generation framework that integrates differentiable optimization over token logits into the decoding loop.<n>$nabla$-Reasoner achieves over 20% accuracy improvement on a challenging mathematical reasoning benchmark.
  arXiv  Detail & Related papers  (2026-03-05T08:42:54Z)
• Towards Anytime-Valid Statistical Watermarking [63.02116925616554]
  We develop the first e-value-based watermarking framework, Anchored E-Watermarking, that unifies optimal sampling with anytime-valid inference.<n>Our framework can significantly enhance sample efficiency, reducing the average token budget required for detection by 13-15% relative to state-of-the-art baselines.
  arXiv  Detail & Related papers  (2026-02-19T18:32:26Z)
• Uncertainty Quantification for Large Language Model Reward Learning under Heterogeneous Human Feedback [8.538830579425147]
  We study estimation and statistical reward models used in aligning large language (LLMs)<n>A key component of LLM alignment is reinforcement learning from human feedback.
  arXiv  Detail & Related papers  (2025-12-02T20:22:25Z)
• ZIP-RC: Optimizing Test-Time Compute via Zero-Overhead Joint Reward-Cost Prediction [57.799425838564]
  We present ZIP-RC, an adaptive inference method that equips models with zero-overhead inference-time predictions of reward and cost.<n> ZIP-RC improves accuracy by up to 12% over majority voting at equal or lower average cost.
  arXiv  Detail & Related papers  (2025-12-01T09:44:31Z)
• LaSeR: Reinforcement Learning with Last-Token Self-Rewarding [54.72617309922891]
  Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a core paradigm for enhancing the reasoning capabilities of Large Language Models (LLMs)<n>Previous practice requires the LLM to sequentially generate solutions and self-verifications using two separate prompt templates, which significantly reduces efficiency.<n>We propose LaSeR (Reinforcement Learning with Last-Token Self-Rewarding), an algorithm that simply augments the original RLVR loss with a MSE loss.
  arXiv  Detail & Related papers  (2025-10-16T17:55:11Z)
• Accelerated Test-Time Scaling with Model-Free Speculative Sampling [58.69141724095398]
  We introduce STAND (STochastic Adaptive N-gram Drafting), a novel model-free speculative decoding approach.<n>We show that STAND reduces inference latency by 60-65% compared to standard autoregressive decoding.<n>As a model-free approach, STAND can be applied to any existing language model without additional training.
  arXiv  Detail & Related papers  (2025-06-05T07:31:18Z)
• Scalable Best-of-N Selection for Large Language Models via Self-Certainty [65.31658824274894]
  Best-of-N selection is a key technique for improving the reasoning performance of Large Language Models.<n>We propose self-certainty, a novel and efficient metric to estimate response quality without requiring external reward models.<n>Our findings establish self-certainty as a practical and efficient way for improving LLM reasoning capabilities.
  arXiv  Detail & Related papers  (2025-02-25T19:08:07Z)
• LEASE: Offline Preference-based Reinforcement Learning with High Sample Efficiency [11.295036269748731]
  This paper proposes a offLine prEference-bAsed RL with high Sample Efficiency (LEASE) algorithm to generate unlabeled preference data.<n>Considering the pretrained reward model may generate incorrect labels for unlabeled data, we design an uncertainty-aware mechanism to ensure the performance of reward model.
  arXiv  Detail & Related papers  (2024-12-30T15:10:57Z)
• Faster WIND: Accelerating Iterative Best-of-$N$ Distillation for LLM Alignment [81.84950252537618]
  This paper reveals a unified game-theoretic connection between iterative BOND and self-play alignment.<n>We establish a novel framework, WIN rate Dominance (WIND), with a series of efficient algorithms for regularized win rate dominance optimization.
  arXiv  Detail & Related papers  (2024-10-28T04:47:39Z)
• Self-Evolutionary Large Language Models through Uncertainty-Enhanced Preference Optimization [9.618391485742968]
  Iterative preference optimization has recently become one of the de-facto training paradigms for large language models (LLMs) We present an uncertainty-enhanced textbfPreference textbfOptimization framework to make the LLM self-evolve with reliable feedback. Our framework substantially alleviates the noisy problem and improves the performance of iterative preference optimization.
  arXiv  Detail & Related papers  (2024-09-17T14:05:58Z)
• Provably Mitigating Overoptimization in RLHF: Your SFT Loss is Implicitly an Adversarial Regularizer [52.09480867526656]
  We identify the source of misalignment as a form of distributional shift and uncertainty in learning human preferences.<n>To mitigate overoptimization, we first propose a theoretical algorithm that chooses the best policy for an adversarially chosen reward model.<n>Using the equivalence between reward models and the corresponding optimal policy, the algorithm features a simple objective that combines a preference optimization loss and a supervised learning loss.
  arXiv  Detail & Related papers  (2024-05-26T05:38:50Z)
• $i$REPO: $i$mplicit Reward Pairwise Difference based Empirical Preference Optimization [12.266207199002604]
  Large Language Models (LLM) can sometimes produce outputs that deviate from human expectations. We propose a novel framework named $i$REPO, which utilizes implicit Reward pairwise difference regression for Empirical Preference Optimization. We show that $i$REPO effectively achieves self-alignment using soft-label, self-generated responses and the logit of empirical AI annotators.
  arXiv  Detail & Related papers  (2024-05-24T05:42:11Z)
• Learning Off-policy with Model-based Intrinsic Motivation For Active Online Exploration [15.463313629574111]
  This paper investigates how to achieve sample-efficient exploration in continuous control tasks. We introduce an RL algorithm that incorporates a predictive model and off-policy learning elements. We derive an intrinsic reward without incurring parameters overhead.
  arXiv  Detail & Related papers  (2024-03-31T11:39:11Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.