Direct Pass@$k$ Optimization for Thought Pruning: Balancing Quality and Diversity in Test-Time Reasoning
arXiv cs.AI · 23 Oct 2026, 03:59 AEST
A new pruning procedure for reasoning models optimizes the balance between quality and diversity in candidate solutions. This method outperforms existing strategies by maximizing the probability of finding a correct answer among selected partial solutions.
Key facts
- 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.
- 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.
- In this paper, we introduce a pruning procedure that addresses this trade-off directly by optimizing pass@$k$ over the selected solution pool.
- On eight open-weight reasoning models, our method outperforms existing quality-based and diversity-based pruning strategies.
Read aloud
A new pruning procedure for reasoning models optimizes the balance between quality and diversity in candidate solutions. This method outperforms existing strategies by maximizing the probability of finding a correct answer among selected partial solutions. 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. 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. In this paper, we introduce a pruning procedure that addresses this trade-off directly by optimizing pass@$k$ over the selected solution pool. On eight open-weight reasoning models, our method outperforms existing quality-based and diversity-based pruning strategies.
Sources in this cluster
- Direct Pass@$k$ Optimization for Thought Pruning: Balancing Quality and Diversity in Test-Time Reasoning
arXiv cs.AI · 23 Oct 2026, 03:59 AEST