# Learning Generative Selection for Best-of-N

arXiv: https://arxiv.org/abs/2602.02143 · HTML: https://mlsanity.com/paper/2602.02143/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Reinforcement learning for reasoning, one of 17 directions in LLM reasoning.
- Builds on 21 of its 27 references that are in the map.
- First posted on arXiv 2 Feb 2026.

## Placed in

- LLM reasoning / [Reinforcement learning for reasoning](https://mlsanity.com/lr/reinforcement-learning-for-reasoning/)

## Builds on (most cited first, 21 of 21)

- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Self-Consistency Improves Chain of Thought Reasoning in Language Models](https://mlsanity.com/paper/2203.11171/) (2022)
- [DAPO: An Open-Source LLM Reinforcement Learning System at Scale](https://mlsanity.com/paper/2503.14476/) (2025)
- [HybridFlow: A Flexible and Efficient RLHF Framework](https://mlsanity.com/paper/2409.19256/) (2024)
- [Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters](https://mlsanity.com/paper/2408.03314/) (2024)
- [Large Language Monkeys: Scaling Inference Compute with Repeated Sampling](https://mlsanity.com/paper/2407.21787/) (2024)
- [Generative Verifiers: Reward Modeling as Next-Token Prediction](https://mlsanity.com/paper/2408.15240/) (2024)
- [A Survey on Test-Time Scaling in Large Language Models: What, How, Where, and How Well?](https://mlsanity.com/paper/2503.24235/) (2025)
- [OpenMathInstruct-2: Accelerating AI for Math with Massive Open-Source Instruction Data](https://mlsanity.com/paper/2410.01560/) (2024)
- [ACECODER: Acing Coder RL via Automated Test-Case Synthesis](https://mlsanity.com/paper/2502.01718/) (2025)
- [AceReason-Nemotron: Advancing Math and Code Reasoning through Reinforcement Learning](https://mlsanity.com/paper/2505.16400/) (2025)
- [AceMath: Advancing Frontier Math Reasoning with Post-Training and Reward Modeling](https://mlsanity.com/paper/2412.15084/) (2024)
- [Critique Fine-Tuning: Learning to Critique is More Effective than Learning to Imitate](https://mlsanity.com/paper/2501.17703/) (2025)
- [Sample, Scrutinize and Scale: Effective Inference-Time Search by Scaling Verification](https://mlsanity.com/paper/2502.01839/) (2025)
- [The Majority is not always right: RL training for solution aggregation](https://mlsanity.com/paper/2509.06870/) (2025)
- [CodeContests+: High-Quality Test Case Generation for Competitive Programming](https://mlsanity.com/paper/2506.05817/) (2025)
- [Learning to Reason Across Parallel Samples for LLM Reasoning](https://mlsanity.com/paper/2506.09014/) (2025)
- [Heimdall: test-time scaling on the generative verification](https://mlsanity.com/paper/2504.10337/) (2025)
- [OpenCodeReasoning-II: A Simple Test Time Scaling Approach via Self-Critique](https://mlsanity.com/paper/2507.09075/) (2025)
- [Learning to Refine: Self-Refinement of Parallel Reasoning in LLMs](https://mlsanity.com/paper/2509.00084/) (2025)
- [Scoring Verifiers: Evaluating Synthetic Verification for Code and Reasoning](https://mlsanity.com/paper/2502.13820/) (2025)

## Built on by (most cited first, 2 of 2)

- [Primal Generation, Dual Judgment: Self-Training from Test-Time Scaling](https://mlsanity.com/paper/2605.11299/) (2026)
- [Selection, Recombination, or a Fresh Solve? A Candidate-Free Control for Single-Pass Test-Time Aggregation](https://mlsanity.com/paper/2608.18379/) (2026)

## Abstract (arXiv)

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.