Direct Density Ratio Optimization: A Statistically Consistent Approach to Aligning Large Language Models
This addresses a critical gap in LLM alignment for safe deployment by providing a statistically consistent method that works regardless of preference structure.
The paper tackles the problem of aligning large language models with human preferences by introducing Direct Density Ratio Optimization (DDRO), which directly estimates density ratios to avoid assumptions about preference models, and it achieves superior performance on major benchmarks.
Aligning large language models (LLMs) with human preferences is crucial for safe deployment, yet existing methods assume specific preference models like Bradley-Terry model. This assumption leads to statistical inconsistency, where more data doesn't guarantee convergence to true human preferences. To address this critical gap, we introduce a novel alignment method Direct Density Ratio Optimization (DDRO). DDRO directly estimates the density ratio between preferred and unpreferred output distributions, circumventing the need for explicit human preference modeling. We theoretically prove that DDRO is statistically consistent, ensuring convergence to the true preferred distribution as the data size grows, regardless of the underlying preference structure. Experiments demonstrate that DDRO achieves superior performance compared to existing methods on many major benchmarks. DDRO unlocks the potential for truly data-driven alignment, paving the way for more reliable and human-aligned LLMs.