GRAD：面向幻觉缓解的图检索自适应解码
GRAD: Graph-Retrieved Adaptive Decoding for Hallucination Mitigation
AI总结：
本文提出解码时方法 GRAD，通过从检索语料构建稀疏 token 转移图并自适应融合 logits，在不重新训练的情况下缓解大语言模型幻觉，在多个问答基准上显著优于贪心解码等基线。
摘要：
即使模型规模不断增大，幻觉缓解仍然是大语言模型（LLMs）面临的一项持久挑战。现有方法通常依赖外部知识源，例如结构化数据库或知识图谱，通过提示或检索来访问。然而，基于提示的 grounding 较为脆弱且对领域敏感，而符号化知识集成则会带来高昂的检索和格式化成本。受知识图谱启发，我们提出了 Graph-Retrieved Adaptive Decoding（GRAD），这是一种解码阶段的方法，无需重新训练即可将生成过程 grounded 在语料库衍生的证据上。GRAD 通过在单个前向传播中累积一个小规模检索语料库上的 next-token logits，构建一个稀疏的 token 转移图。在解码过程中，图检索得到的 logits 经过 max-normalization 后与模型 logits 自适应融合，以倾向于高证据支持的续写，同时保持流畅性。在三个模型以及涵盖内在幻觉、外在幻觉和事实性任务的一系列问答基准上，GRAD 持续超越基线方法，与贪心解码相比，内在准确率最高提升 9.7%，幻觉率降低 8.6%，正确性提升 6.9%，并在所有方法中取得了最高的 truth–informativeness product 分数。GRAD 为对比解码和知识图谱增强提供了一种轻量级、即插即用的替代方案，表明来自语料库级 token 转移的统计证据能够有效引导生成朝向更真实且可验证的输出。
英文摘要：
Hallucination mitigation remains a persistent challenge for large language models (LLMs), even as model scales grow. Existing approaches often rely on external knowledge sources, such as structured databases or knowledge graphs, accessed through prompting or retrieval. However, prompt-based grounding is fragile and domain-sensitive, while symbolic knowledge integration incurs heavy retrieval and formatting costs. Motivated by knowledge graphs, we introduce Graph-Retrieved Adaptive Decoding (GRAD), a decoding-time method that grounds generation in corpus-derived evidence without retraining. GRAD constructs a sparse token transition graph by accumulating next-token logits across a small retrieved corpus in a single forward pass. During decoding, graph-retrieved logits are max-normalized and adaptively fused with model logits to favor high-evidence continuations while preserving fluency. Across three models and a range of question-answering benchmarks spanning intrinsic, extrinsic hallucination, and factuality tasks, GRAD consistently surpasses baselines, achieving up to 9.7$\%$ higher intrinsic accuracy, 8.6$\%$ lower hallucination rates, and 6.9$\%$ greater correctness compared to greedy decoding, while attaining the highest truth--informativeness product score among all methods. GRAD offers a lightweight, plug-and-play alternative to contrastive decoding and knowledge graph augmentation, demonstrating that statistical evidence from corpus-level token transitions can effectively steer generation toward more truthful and verifiable outputs.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
