Newest papers

Direct Pass@$k$ Optimization for Thought Pruning: Balancing Quality and Diversity in Test-Time Reasoning

Mahdi Salmani, Adel Javanmard, Devansh Gupta, Meisam Razaviyayn · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.31667 · cited by 0 works (OpenAlex) · open-access PDF

Russian

Сохраняя все решения кандидатов до завершения может быть вычислительно требовательным, так что недавняя работа изготавливает балл кандидатов после создания только короткого префикса, совершая дальнейшее вычисление только для выживших кандидатов. Существующие стратегии изготавливания широко делятся на две категории: методы на основе качества одновременно назначают балл каждому частичному решению и отбрасывают предсказанные, чтобы в конечном итоге привести к неправильному окончательному решению, в то время как методы, основанные на разнообразии, выбирают частичные решения, которые следуют за разными траекториями вычисления и поэтому, вероятно, будут покрывать различные окончательные решения. Напряжение между этими двумя целями является примером пенопластной эксплуатации, но ни малейшей работой

Machine translation (Meta M2M100); the English original follows.

Abstract

Test-time scaling has emerged as a key inference strategy for reasoning models, where a model generates a large number of candidate solutions in the hope of finding at least one correct solution. Running all candidate solutions to completion can be computationally demanding, so recent work prunes the candidate pool after generating only a short prefix, committing further computation only to the surviving candidates. Existing pruning strategies divide broadly into two categories: Quality-based methods assign a score to each partial solution and discard those predicted to ultimately lead to an incorrect final solution, while diversity-based methods select partial solutions that follow different reasoning trajectories and are therefore likely to cover different final solutions. The tension between these two objectives is an instance of exploration versus exploitation, yet little to no prior work on pruning exposes this trade-off directly, let alone optimizes it. In this paper, we introduce a pruning procedure that addresses this trade-off directly by optimizing pass@$k$ over the selected solution pool. In particular, it chooses the pool to maximize the probability that at least one of its partial solutions, when continued to completion, yields a correct answer. Because this objective simultaneously rewards high-quality partial solutions and penalizes redundancy among them, quality and diversity are naturally balanced by the criterion itself rather than through separate ad-hoc design choices. On eight open-weight reasoning models, our method outperforms existing quality-based and diversity-based pruning strategies. The gains are particularly pronounced when the pruning procedure keeps a moderate number of partial solutions for continued generation, highlighting the value of directly balancing quality and diversity during pruning.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Pruning, Computer science, Artificial intelligence, Redundancy (engineering), Inference, Machine learning

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange