Adaptive Blockwise Search: Inference-Time Alignment for Large Language Models
On the map: LLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Abstract (arXiv)
LLM alignment remains a critical challenge. Inference-time methods provide a flexible alternative to fine-tuning, but their uniform computational effort often yields suboptimal alignment. We hypothesize that for many alignment tasks, the initial tokens of a response are disproportionately more critical. To leverage this principle, we introduce AdaSearch, a novel blockwise search strategy. It adaptively allocates a fixed computational budget using a sampling schedule, focusing search effort on these critical tokens. We apply AdaSearch to sequential decoding and introduce its tree-search counterpart, AdaBeam. Our comprehensive evaluation across eight LLMs demonstrates that AdaSearch outperforms strong Best-of-N and fine-tuning baselines. Specifically, win-rates improve by over 10% for harmlessness generation, controlled sentiment generation, and for mathematical reasoning tasks relative to Best-of-N.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in LLM preference optimization methods, one of 10 directions in LLM alignment / RLHF / preference optimisation. Builds on 24 of its 31 references that are in the map. First posted on arXiv 27 Oct 2025.
Builds on
24 of 31 references.
Closest ancestors · 6
Same direction, newest first
- 2025Evaluation of Best-of-N Sampling Strategies for Language Model Alignment18 Feb
- 2024TreeBoN: Enhancing Inference-Time Alignment with Speculative Tree-Search and Best-of-N Sampling18 Oct
- 2024Cascade Reward Sampling for Efficient Decoding-Time Alignment24 Jun
- 2024ARGS: Alignment as Reward-Guided Search23 Jan
- 2023RAIN: Your Language Models Can Align Themselves without Finetuning13 Sep
- 2022Training language models to follow instructions with human feedback4 Mar
Other references · 18
- 2021Training Verifiers to Solve Math Word Problemscited by 9,743
- 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
- 2021Finetuned Language Models Are Zero-Shot Learnerscited by 3,747
- 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,634
- 2023Let's Verify Step by Stepcited by 3,598
- 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Modelscited by 3,155
- 2020Learning to summarize from human feedbackcited by 2,562
- 2020DeBERTa: Decoding-enhanced BERT with Disentangled Attentioncited by 2,157
- 2021WebGPT: Browser-assisted question-answering with human feedbackcited by 1,722
- 2022Scaling Laws for Reward Model Overoptimizationcited by 1,056
- 2023Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotationscited by 895
- 2022Solving math word problems with process- and outcome-based feedbackcited by 871
6 more
- 2021FUDGE: Controlled Text Generation With Future Discriminatorscited by 374
- 2023AI Alignment: A Comprehensive Surveycited by 256
- 2023Controlled Decoding from Language Modelscited by 133
- 2024Fast Best-of-N Decoding via Speculative Rejectioncited by 112
- 2023Reward-Augmented Decoding: Efficient Controlled Text Generation With a Unidirectional Reward Modelcited by 79
- 2025Accelerating Large Language Model Reasoning via Speculative Searchcited by 16
Built on by
Citing papers per year: 2026: 1.
1 paper.
Similar, not cited
Close in wording, no citation either way
- 2024Inference Time Alignment with Reward-Guided Tree Search
- 2026Inference-Aware Meta-Alignment of LLMs via Non-Linear GRPO
- 2025AdaBoN: Adaptive Best-of-N Alignment
- 2025Alignment-Aware Decoding
- 2024Nudging: Inference-time Alignment of LLMs via Guided Decoding
- 2025On Almost Surely Safe Alignment of Large Language Models at Inference-Time
- 2024Efficient Alignment of Large Language Models via Data Sampling
- 2024Inference-Time Language Model Alignment via Integrated Value Guidance