Learning Generative Selection for Best-of-N

• URL: http://arxiv.org/abs/2602.02143v1
• Date: Mon, 02 Feb 2026 14:21:15 GMT
• Title: Learning Generative Selection for Best-of-N
• Authors: Shubham Toshniwal, Aleksander Ficek, Siddhartha Jain, Wei Du, Vahid Noroozi, Sadegh Mahdavi, Somshubra Majumdar, Igor Gitman,
• Abstract summary: We show that small reasoning models can acquire strong GenSelect capabilities through targeted reinforcement learning.<n>Our results establish reinforcement learning as a scalable way to unlock strong generative selection in small models.
• Score: 52.88943295436412
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Scaling test-time compute via parallel sampling can substantially improve LLM reasoning, but is often limited by Best-of-N selection quality. Generative selection methods, such as GenSelect, address this bottleneck, yet strong selection performance remains largely limited to large models. We show that small reasoning models can acquire strong GenSelect capabilities through targeted reinforcement learning. To this end, we synthesize selection tasks from large-scale math and code instruction datasets by filtering to instances with both correct and incorrect candidate solutions, and train 1.7B-parameter models with DAPO to reward correct selections. Across math (AIME24, AIME25, HMMT25) and code (LiveCodeBench) reasoning benchmarks, our models consistently outperform prompting and majority-voting baselines, often approaching or exceeding much larger models. Moreover, these gains generalize to selecting outputs from stronger models despite training only on outputs from weaker models. Overall, our results establish reinforcement learning as a scalable way to unlock strong generative selection in small models, enabling efficient test-time scaling.

Related papers

• Satori-SWE: Evolutionary Test-Time Scaling for Sample-Efficient Software Engineering [51.7496756448709]
  Language models (LMs) perform well on coding benchmarks but struggle with real-world software engineering tasks.<n>Existing approaches rely on supervised fine-tuning with high-quality data, which is expensive to curate at scale.<n>We propose Test-Time Scaling (EvoScale), a sample-efficient method that treats generation as an evolutionary process.
  arXiv  Detail & Related papers  (2025-05-29T16:15:36Z)
• Scalable Best-of-N Selection for Large Language Models via Self-Certainty [75.1351701045874]
  Best-of-N selection is a key technique for improving the reasoning performance of Large Language Models (LLMs)<n>We propose self-certainty, a novel and efficient metric that leverages the inherent probability distribution of LLM outputs 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)
• Efficient Response Generation Strategy Selection for Fine-Tuning Large Language Models Through Self-Aligned Perplexity [29.665161650753742]
  Fine-tuning large language models (LLMs) typically relies on producing large sets of input-output pairs.<n>Recent findings show that how these training outputs are generated can significantly affect the performance of the fine-tuned model.<n>This paper proposes a scalable approximate method that assesses a small subset of generated data to estimate its suitability for a specific target LLM.
  arXiv  Detail & Related papers  (2025-02-17T13:14:11Z)
• How to Select Pre-Trained Code Models for Reuse? A Learning Perspective [29.83900262273832]
  Off-the-shelf Pre-trained Code Models (PCMs) have been released publicly.<n>We conduct experiments on 100 widely-used opensource PCMs for code intelligence tasks, with sizes ranging from 42.5 million to 3 billion parameters.<n>The results demonstrate that learning-based selection methods reduce selection time to 100 seconds, compared to 2,700 hours with brute-force fine-tuning.
  arXiv  Detail & Related papers  (2025-01-07T13:45:24Z)
• A Two-Phase Recall-and-Select Framework for Fast Model Selection [13.385915962994806]
  We propose a two-phase (coarse-recall and fine-selection) model selection framework. It aims to enhance the efficiency of selecting a robust model by leveraging the models' training performances on benchmark datasets. It has been demonstrated that the proposed methodology facilitates the selection of a high-performing model at a rate about 3x times faster than conventional baseline methods.
  arXiv  Detail & Related papers  (2024-03-28T14:44:44Z)
• DsDm: Model-Aware Dataset Selection with Datamodels [81.01744199870043]
  Standard practice is to filter for examples that match human notions of data quality. We find that selecting according to similarity with "high quality" data sources may not increase (and can even hurt) performance compared to randomly selecting data. Our framework avoids handpicked notions of data quality, and instead models explicitly how the learning process uses train datapoints to predict on the target tasks.
  arXiv  Detail & Related papers  (2024-01-23T17:22:00Z)
• GistScore: Learning Better Representations for In-Context Example Selection with Gist Bottlenecks [3.9638110494107095]
  In-context Learning (ICL) is the ability of Large Language Models (LLMs) to perform new tasks when conditioned on prompts. We propose Example Gisting, a novel approach for training example encoders through supervised fine-tuning. We show that our fine-tuned models get state-of-the-art ICL performance with over 20% absolute gain over off-the-shelf retrievers.
  arXiv  Detail & Related papers  (2023-11-16T06:28:05Z)
• MILO: Model-Agnostic Subset Selection Framework for Efficient Model Training and Tuning [68.12870241637636]
  We propose MILO, a model-agnostic subset selection framework that decouples the subset selection from model training. Our empirical results indicate that MILO can train models $3times - 10 times$ faster and tune hyperparameters $20times - 75 times$ faster than full-dataset training or tuning without performance.
  arXiv  Detail & Related papers  (2023-01-30T20:59:30Z)
• True Few-Shot Learning with Language Models [78.42578316883271]
  We evaluate the few-shot ability of LMs when held-out examples are unavailable. Our findings suggest that prior work significantly overestimated the true few-shot ability of LMs.
  arXiv  Detail & Related papers  (2021-05-24T17:55:51Z)

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.