Hallucination Self-Play: Bootstrapping Reinforced Detector via Evolved Generator
For practitioners needing to detect hallucinations in LLM outputs, HSP offers an iterative self-improvement framework that reduces reliance on expensive human annotations.
Hallucination Self-Play (HSP) bootstraps a hallucination detector by using it as a reward model to train a generator via RLAIF, which then produces harder-to-detect hallucinations to further improve the detector. On RAGTruth, a small LLM matches or outperforms advanced LLMs without external supervision.
Identifying faithfulness hallucinations in LLM-generated outputs remains challenging due to the scarcity of high-quality annotated data. Recent work relies on advanced LLMs to synthesize training data, including rationales, labels, and hallucinated claims. However, these methods treat the generator as a static component, limiting iterative improvement of the detector. To address this limitation, we introduce Hallucination Self-Play (HSP), a novel framework that enables the detector to bootstrap with an evolved generator. HSP involves two roles initialized from the same base model, a detector that assesses the faithfulness of model outputs, and a generator that produces increasingly hard-to-detect hallucinated responses. Specifically, the detector is first fine-tuned on human-labeled data and then employed as a reward model to train the generator via reinforcement learning from AI feedback (RLAIF). In turn, the evolved generator synthesizes hallucination data to further optimize the detector through rule-based reinforcement learning. Experiments on RAGTruth benchmark and two model families demonstrate that the proposed framework can progressively enhance a small LLM to match or even outperform advanced LLMs without external supervision. Our code is available at https://anonymous.4open.science/r/Hallucination-Self-Play-50B5 .