arXiv · cs.AI · Suwan Wu, Yumeng Lin, Pengcheng Yuan, Xiaolong Jiang · 10 小时前模型评分42
SIRF：面向工业内容风控的规范内化风险基础模型
SIRF: A Spec-Internalized Risk Foundation Model for Industrial Content Risk Control
导读
SIRF 将平台复杂策略内化进模型权重，在仅输出判定结果的低延迟部署下实现高精度风控。同源对比中，SIRF-8B-受监督微调（SFT，Supervised Fine-Tuning） 的 Black Recall@P95 达 71.3%，比基线高 15.1 个百分点，仅用约 70M CPT token 且不损害通用能力。
分类置信度
中 · 待人工复核
- 只有单一媒体来源，需要人工复核。
来源级别：媒体与个人
事件类型：研究发布
证据摘录
- “We present SIRF (Spec-Internalized Risk Foundation Model), which internalizes a platform's complex policies, synthesized without additional human annotation via EntiGraph, MAGA rewriting and account-level chain-of-thought (CoT), into the weights via continued pretraining (CPT), so rules are applied at high precision under an ultra-low-latency, verdict-only deployment.”
来源：arXiv · cs.AI · arxiv.org