Self-play preference optimization has emerged as a prominent paradigm for aligning large language models (LLMs). It typically involves a language model to generate on-policy responses for prompts and a reward model (RM) to guide the selection of chosen and rejected responses, which can be further trained with direct preference optimization (DPO). However, the role of prompts remains underexplored, despite being a core component in this pipeline. In this work, we investigate how prompts of varying difficulty influence self-play preference optimization. We use the mean reward of sampled responses of a prompt as a proxy for its difficulty. We first find that difficult prompts exhibit substantially inferior self-play optimization performance compared to easy prompts for language models. Moreover, incorporating difficult prompts into training fails to enhance overall performance and, in fact, leads to slight degradation compared to training on easy prompts alone. Third, there is a clear upward trend in optimization performance as prompt difficulty decreases. We also observe that the performance gap between difficult and easy prompts tends to close as the model capacity increases, suggesting that prompt difficulty interacts with the model capacity. Building on these findings, we explore strategies to mitigate the adversary effect of difficult prompts on final performance. We demonstrate that only training on a small portion (30%) of the easiest prompts improves overall self-play performance on AlpacaEval 2 and Arena-Hard. We also report failed attempts and lessons learned.

Iterative data generation and model retraining are widely used to align large language models (LLMs).It typically involves a policy model to generate on-policy responses and a reward model to guide training data selection. Direct Preference Optimization (DPO) further enhances this process by constructing preference pairs of chosen and rejected responses. In this work, we aim to scale up the number of on-policy samples via repeated random sampling to improve alignment performance. Conventional practice selects the sample with the highest reward as chosen and the lowest as rejected for DPO. However, our experiments reveal that this strategy leads to a decline in performance as the sample size increases. To address this, we investigate preference data construction through the lens of underlying normal distribution of sample rewards. We categorize the reward space into seven representative points and systematically explore all 21 (C₇²) pairwise combinations. Through evaluations on four models using AlpacaEval 2, we find that selecting the rejected response at reward position 𝜇 - 2𝜎 rather than the minimum reward, is crucial for optimal performance. We finally introduce a scalable preference data construction strategy that consistently enhances model performance as the sample scale increases.