GCPO: When Contrast Fails, Go Gold
Hao Wu
Abstract
Group Contrastive Policy Optimization (GCPO) enhances reinforcement learning for large language models by incorporating external reference answers, improving training efficiency and generalization.
Reinforcement learning has been widely applied to enhance the reasoning capabilities of large language models. Extending the inference limits of smaller models has become a prominent research focus. However, algorithms such as Group Relative Policy Optimization (GRPO) suffer from a clear drawback: the upper bound of a model's rollout responses is entirely determined by the model itself, preventing the acquisition of knowledge from samples that are either all incorrect or all correct. In this paper, we introduce Group Contrastive Policy Optimization (GCPO), a method that incorporates external standard reference answers. When the model cannot solve a problem, the reference answer supplies the correct response, steering the model toward an unequivocally accurate update direction. This approach offers two main advantages: (1) it improves training efficiency by fully utilizing every sample; (2) it enables the model to emulate the problem solving strategy of the reference answer during training, thereby enhancing generalization in reasoning. GCPO achieves outstanding results across multiple benchmark datasets, yielding substantial improvements over the baseline model. Our code is available at: https://github.com/AchoWu/GCPO.
Community
GCPO (Group Contrastive Policy Optimization) is a novel reinforcement learning algorithm designed to enhance the reasoning capabilities of language models, especially in scenarios where the model fails to generate correct responses. Unlike previous methods like GRPO, which rely solely on the model’s own rollouts, GCPO introduces Golden Answers (GAs) — external reference answers — to guide the model’s updates when all sampled responses are incorrect.
GCPO (Group Contrastive Policy Optimization) is a novel reinforcement learning algorithm designed to enhance the reasoning capabilities of language models, especially in scenarios where the model fails to generate correct responses. Unlike previous methods like GRPO, which rely solely on the model’s own rollouts, GCPO introduces Golden Answers (GAs) — external reference answers — to guide the model’s updates when all sampled responses are incorrect.
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
The following papers were recommended by the Semantic Scholar API
- ConfClip: Confidence-Weighted and Clipped Reward for Reinforcement Learning in LLMs (2025)
- Unlocking Reasoning Capabilities in LLMs via Reinforcement Learning Exploration (2025)
- ToolExpander: Extending the Frontiers of Tool-Using Reinforcement Learning to Weak LLMs (2025)
- Improving Sampling Efficiency in RLVR through Adaptive Rollout and Response Reuse (2025)
- Your Models Have Thought Enough: Training Large Reasoning Models to Stop Overthinking (2025)
- More Than One Teacher: Adaptive Multi-Guidance Policy Optimization for Diverse Exploration (2025)
- Beyond Pass@1: Self-Play with Variational Problem Synthesis Sustains RLVR (2025)
Please give a thumbs up to this comment if you found it helpful!
If you want recommendations for any Paper on Hugging Face checkout this Space
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment:
@librarian-bot
recommend
Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper
Collections including this paper 0
No Collection including this paper