CPPO: Accelerating the Training of Group Relative Policy Optimization-Based Reasoning Models

Jun-23-2026, 07:45:01 GMT–Neural Information Processing Systems

This paper introduces Completion Pruning Policy Optimization (CPPO) to accelerate the training of reasoning models based on Group Relative Policy Optimization (GRPO).

completion, large language model, machine learning, (19 more...)

Neural Information Processing Systems

Jun-23-2026, 07:45:01 GMT

Conferences PDF

Add feedback

Country:
- Asia > China (0.46)

Genre:
- Research Report
  - Experimental Study (1.00)
  - New Finding (0.68)

Industry:
- Education (0.46)

Technology:
- Information Technology > Artificial Intelligence
  - Natural Language > Large Language Model (0.95)
  - Representation & Reasoning (0.94)
  - Machine Learning > Neural Networks
    - Deep Learning (0.68)