Code Consistency Preference Optimization Verification for Language Model Alignment

arXiv:2609.19002v1 Announce Type: cross Abstract: Execution-based verification enhances large language models' mathematical reasoning through computational soundness and dependency-aware filtering. However, prior preference optimization methods relying on Bradley-Terry reward models fail to capture the logical dependencies and execution consistency needed for scientific tasks. We propose a method that generates computationally sound solutions with dependency graphs for execution-consistent preference optimization. We first build a scientific reasoning dataset using UltraFeedback prompts, model generations, verification, and consistency results. Then we extract reasoning step expressions, prerequisites, and derivability relationships to construct dependency graphs and compute execution consistency scores. These scores are appended to each step, creating paired training data. Fine-tuning Llama-3-8B and DeepSeekMath-7B yields significant gains: +17.0% on MATH and +15.1% on GSM8K. Extending our Scientific Feasibility Control framework achieves 50.1% accuracy on PhyX multimodal physics reasoning, surpassing DeepSeek-R1 (49.8%) and OpenAI o3-mini (48.2%), with 91.7% scientific validity coverage at alpha=0.10 and 73% fewer scientific law violations, resulting in the CCPO model family.

Abstract:Execution-based verification enhances large language models' mathematical reasoning through computational soundness and dependency-aware filtering. However, prior preference optimization methods relying on Bradley-Terry reward models fail to capture the logical dependencies and execution consistency needed for scientific tasks. We propose a method that generates computationally sound solutions with dependency graphs for execution-consistent preference optimization. We first build a scientific reasoning dataset using UltraFeedback prompts, model generations, verification, and consistency results. Then we extract reasoning step expressions, prerequisites, and derivability relationships to construct dependency graphs and compute execution consistency scores. These scores are appended to each step, creating paired training data. Fine-tuning Llama-3-8B and DeepSeekMath-7B yields significant gains: +17.0% on MATH and +15.1% on GSM8K. Extending our Scientific Feasibility Control framework achieves 50.1% accuracy on PhyX multimodal physics reasoning, surpassing DeepSeek-R1 (49.8%) and OpenAI o3-mini (48.2%), with 91.7% scientific validity coverage at alpha=0.10 and 73% fewer scientific law violations, resulting in the CCPO model family.

Software Engineering (cs.SE); Computation and Language (cs.CL)

arXiv:2609.19002 [cs.SE]

(or arXiv:2609.19002v1 [cs.SE] for this version)

https://doi.org/10.48550/arXiv.2609.19002 Focus to learn more arXiv-issued DOI via DataCite (pending registration)

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

analiza aiqbench

Code Consistency Preference Optimization Verification for Language Model Alignment

Subiecte conexe