Answer First, Reason Later: Aligning Search Relevance via Mode-Balanced Reinforcement Learning

Building a search relevance model that achieves both low latency and high performance is a long-standing challenge in the search industry. To satisfy the millisecond-level response requirements of online systems while retaining the interpretable reasoning traces of Large Language Models (LLMs), we propose a novel Answer-First, Reason Later (AFRL) paradigm. This paradigm requires the model to output the definitive relevance score in the very first token, followed by a structured logical explanation. Inspired by the success of reasoning models, we adopt a "Supervised Fine-Tuning (SFT) + Reinforcement Learning (RL)" pipeline to achieve AFRL. However, directly applying existing RL training often leads to mode collapse in the search relevance task, where the model forgets complex long-tail rules in pursuit of high rewards. From an information theory perspective: RL inherently minimizes the Reverse KL divergence, which tends to seek probability peaks (mode-seeking) and is prone to "reward hacking." On the other hand, SFT minimizes the Forward KL divergence, forcing the model to cover the data distribution (mode-covering) and effectively anchoring expert rules. Based on this insight, we propose a Mode-Balanced Optimization strategy, incorporating an SFT auxiliary loss into Stepwise-GRPO training to balance these two properties. Furthermore, we construct an automated instruction evolution system and a multi-stage curriculum to ensure expert-level data quality. Extensive experiments demonstrate that our 32B teacher model achieves state-of-the-art performance. Moreover, the AFRL architecture enables efficient knowledge distillation, successfully transferring expert-level logic to a 0.6B model, thereby reconciling reasoning depth with deployment latency.