十亿以下，超级前沿：小语言模型在通用和文学关系抽取上媲美零样本前沿LLM
Sub-Billion, Super-Frontier: Small Language Models Rival Zero-Shot Frontier LLMs on General and Literary Relation Extraction
AI总结：
研究小语言模型（SLM）在通用和文学关系抽取任务上能否通过任务适配缩小与零样本前沿LLM的差距，发现4位SLM经微调后性能超越GPT-5.4和Claude Sonnet 4.6，归因于任务适配而非生成解码。
摘要：
大型语言模型（LLM）在关系抽取（RE）上表现强劲，但其计算需求和依赖专有API限制了在资源受限或隐私敏感场景下的部署。我们研究小语言模型（SLM）能在多大程度上缩小这一差距，涵盖通用领域和文学文本。我们在三种领域组合机制和两种提示条件微调风格（30种配置）下评估了从360M到3B参数的五个模型，并与零样本前沿LLM和判别式RoBERTa基线进行比较。在九个基准上，最佳十亿以下模型Qwen2.5-0.5B在混合通用领域数据上微调后，通用领域正类微F1达到0.83，而GPT-5.4为0.69，Claude Sonnet 4.6为零样本评估为0.66。这并不意味着SLM本质上更强；相反，针对性的任务适配使得可在单个消费级GPU上部署的4位模型在此协议下超越通用前沿系统。领域内RoBERTa基线也超过两个前沿模型，表明增益来自任务适配而非生成解码。在文学RE上，微调后的SLM在人工标注的Biographical基准上达到0.92（GPT-5.4为0.83），在两个基准的文学平均上达到0.833（GPT-5.4为0.578）。针对性的领域自适应预训练案例研究相比监督微调未产生实际有意义的增益，而最干净的族内规模比较仅显示边际改进。这些结果表明，当任务特定数据可用时，紧凑的任务适配模型可以提供准确、隐私且硬件高效的RE。
英文摘要：
Large language models (LLMs) achieve strong relation extraction (RE), but their computational demands and reliance on proprietary APIs limit deployment in resource-constrained or privacy-sensitive settings. We investigate how far small language models (SLMs) can close this gap across general-domain and literary text. We evaluate five models from 360M to 3B parameters under three domain-composition regimes and two prompt-conditioned tuning styles (30 configurations), comparing them with zero-shot frontier LLMs and a discriminative RoBERTa baseline. Across nine benchmarks, the best sub-billion model, Qwen2.5-0.5B fine-tuned on pooled general-domain data, achieves a general-domain positive-class micro-F1 of 0.83, versus 0.69 for GPT-5.4 and 0.66 for Claude Sonnet 4.6 evaluated zero-shot. This does not imply that SLMs are intrinsically stronger; rather, targeted task adaptation enables 4-bit models deployable on a single consumer GPU to outperform general-purpose frontier systems under this protocol. An in-domain RoBERTa baseline also exceeds both frontier models, indicating that the gain stems from task adaptation rather than generative decoding. On literary RE, tuned SLMs reach 0.92 on the human-annotated Biographical benchmark versus 0.83 for GPT-5.4, and 0.833 versus 0.578 on the two-benchmark literary average. A targeted domain-adaptive pretraining case study yields no practically meaningful gain over supervised fine-tuning, while the cleanest within-family scale comparison shows only marginal improvement. These results show that, when task-specific data are available, compact task-adapted models can provide accurate, private, and hardware-efficient RE.
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。
