Direct Pass@$k$ Optimization for Thought Pruning: Balancing Quality and Diversity in Test-Time Reasoning
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 comput
Reading in Plain English mode — toggle below to switch
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 comput
This paper addresses a critical challenge in scaling reasoning models by optimizing test-time inference. For ML engineers, understanding these pruning strategies is key to deploying computationally efficient and effective large language models, especially in scenarios requiring complex multi-step reasoning.