Data for Agents 智能体数据
- +10
Image: Nemotron Post-Training v3 Prompt Atlas 图片：Nemotron 训练后提示词图谱
More Than Model Weights 远不止模型权重
Building AI agents is hard, because the real world does not behave like a benchmark.构建 AI 智能体困难重重，因为现实世界并不像基准测试那样按部就班。
An agent that can't recover from a broken API call, or a workflow it has never seen, is not really an agent. It is an autocompleter with tools. Getting from one to the other is a data problem: software engineering traces, tool-use failures, multi-step reasoning, retrieval, safety, user simulation, workflow execution, and eventually physical world interaction. That is where NVIDIA Nemotron's open data products live.如果一个智能体无法从中断的 API 调用或从未见过的流程中恢复，那么它就称不上真正的智能体，仅仅是一个带有工具的自动补全工具。要实现从前者到后者的跨越，本质上是一个数据问题：涵盖软件工程轨迹、工具使用失败、多步推理、检索、安全性、用户模拟、工作流执行，以及最终的物理世界交互。这正是 NVIDIA Nemotron 开放数据产品所致力于解决的领域。
NVIDIA recently highlighted how open models are driving AI research and showing up across the popular International Conference on Machine Learning (ICML), with nearly 145 papers citing Nemotron models and datasets. Synthetic data plays an important role across that ecosystem:NVIDIA 最近强调了开放模型如何推动 AI 研究，并在国际机器学习会议（ICML）等主流会议上大放异彩，近 145 篇论文引用了 Nemotron 模型和数据集。合成数据在这一生态系统中发挥着重要作用：
- Nemotron-CC used synthetics to enhance the popular Common Crawl dataset for pretraining.Nemotron-CC 利用合成数据增强了用于预训练的常用 Common Crawl 数据集。
- Nemotron-CC-MATH leverages synthetic math questions to improve reasoning.Nemotron-CC-MATH 利用合成数学题来提升推理能力。
- Nemotron Pretraining is a broad collection spanning general, code, math, and synthetic data across trillions of tokens.Nemotron 预训练数据集是一个涵盖通用、代码、数学和合成数据的庞大集合，包含数万亿个 Token。
Part of why NVIDIA releases open datasets is to learn with the community to expand upon these various applications.NVIDIA 发布开放数据集的部分原因，是希望与社区共同学习，拓展这些数据集的各种应用场景。
Open weights matter. But for agents, weights are only part of the story. Reproducibility also depends on the datasets, curation choices, training recipes, and evaluation methods behind the model.开放权重固然重要，但对于智能体而言，权重只是冰山一角。可复现性还取决于模型背后的数据集、筛选标准、训练配方和评估方法。
Agent behavior needs to be inspectable. If a model calls tools, executes workflows, retrieves information, and acts across systems, developers need to understand the data that shaped those behaviors. Open data makes agent behavior inspectable and explainable. Synthetic data is a key piece of the puzzle to making that possible.智能体的行为必须是可审查的。如果模型调用工具、执行工作流、检索信息并跨系统操作，开发者就需要了解塑造这些行为的数据。开放数据使智能体行为变得可审查、可解释，而合成数据正是实现这一目标的关键拼图。
Keep It Like a Secret 保守秘密
NVIDIA's VP of Applied Deep Learning Research Bryan Catanzaro recently noted: "every company is built around a secret" — a workflow, corpus, or customer pattern competitors don't have. Those secrets make AI useful, but companies shouldn't casually expose them. Synthetic data gives teams a way to preserve useful signals without exposing the underlying sources.NVIDIA 应用深度学习研究副总裁 Bryan Catanzaro 最近指出：“每家公司都围绕着一个秘密建立”——即竞争对手所不具备的工作流、语料库或客户模式。这些秘密使 AI 具有实用价值，但公司不应随意公开它们。合成数据为团队提供了一种既能保留有用信号，又不暴露底层来源的方法。
Bryan also talks about cultivating a diverse and participatory AI ecosystem where many kinds of companies, researchers, governments, and communities can contribute. That is not just a value claim. It is a data claim.Bryan 还谈到了培育一个多元且具有参与性的 AI 生态系统，让各类公司、研究人员、政府和社区都能做出贡献。这不仅是一个价值观主张，更是一个数据主张。
If every model learns from the same narrow pool of data, we should not be surprised when the models start to feel the same. The hard part is that the most useful data often sits inside organizations that cannot or will not publish it directly. Everyone benefits from a richer shared data layer. No one wants to be the first to give away the thing that makes them special.如果每个模型都从同一个狭窄的数据池中学习，那么模型变得千篇一律也就不足为奇了。难点在于，最有价值的数据往往掌握在那些无法或不愿直接发布数据的机构手中。共享一个更丰富的数据层对所有人都有利，但没人愿意率先交出让自己与众不同的核心资产。
Synthetic data, released openly, is one way to change that math.公开释放合成数据，是改变这种现状的一种方式。
Exploring Agent Data 探索智能体数据
As part of Nemotron open data, we've released over 10 trillion pre-training tokens and millions of post-training samples spanning many domains and data shapes. That's a lot to make sense of — and raw dataset tables don't help much.作为 Nemotron 开放数据的一部分，我们发布了超过 10 万亿个预训练 Token 和数百万个训练后样本，涵盖了多种领域和数据形态。要理清这些数据并不容易，仅仅依靠原始数据集表格作用有限。
To make it easier to explore what's actually in Nemotron post-training data, we built the Nemotron Post-Training v3 Prompt Atlas: an interactive visual map where each point is a prompt sample, drawn from the Nemotron v3 post-training collection and volume-sampled to reflect the honest proportions of the data mixture.为了更轻松地探索 Nemotron 训练后数据的内容，我们构建了 Nemotron 训练后 v3 提示词图谱（Prompt Atlas）：这是一个交互式可视化地图，其中每个点都是一个提示词样本，均取自 Nemotron v3 训练后数据集，并经过采样以反映数据混合的真实比例。
Color overlays and filters let you reorganize the map by dataset, pipeline stage, domain, or tool use. Since semantically similar prompts cluster together, you can zoom into a region — coding algorithms, safety, math, agentic behavior — inspect representative examples, and use that signal to curate data, build evals, or understand why a model behaves the way it does.通过颜色叠加和过滤器，您可以按数据集、流水线阶段、领域或工具使用情况重新组织地图。由于语义相似的提示词会聚类在一起，您可以缩放至特定区域（如编码算法、安全性、数学、智能体行为），查看代表性示例，并利用这些信号来筛选数据、构建评估体系或理解模型行为背后的原因。
Viva La Persona 万岁，个性化
Agents also need to understand people they are built to support, and this is where “data quality” becomes local, not universal. A toxicity classifier trained on English internet data can miss hostile messages in Korean or Japanese, where aggression is often encoded in politeness levels rather than obvious vocabulary. Same signal, different context. Teams are already grounding agents this way. 智能体还需要理解它们所服务的对象，而这正是“数据质量”变得因地制宜而非普适的地方。一个基于英语互联网数据训练的毒性分类器，可能会漏掉韩语或日语中的敌意信息，因为在这些语言中，攻击性往往隐藏在礼貌用语的层级中，而非明显的词汇里。同样的信号，不同的语境。团队目前正在以这种方式对智能体进行落地调优。
Nemotron-Personas is one attempt at addressing that: locally grounded synthetic personas capturing the diversity and complexity of populations. Built using NeMo Data Designer, NVIDIA’s state-of-the-art compound-AI tooling for synthetic data generation, Nemotron-Personas mirrors official regional demographic and geographic statistics. The goal is not to recreate real people. In a way, it’s to help developers test whether their systems reflect the users, languages, regions, and occupations they claim to serve. Privasis, a derivative dataset built on Nemotron-Personas-USA, shows one direction this could go, layering privacy-preserving synthetic records across medical, financial, legal, and social contexts. Nemotron-Personas 正是解决这一问题的一次尝试：通过本地化的合成角色来捕捉人群的多样性和复杂性。Nemotron-Personas 使用 NVIDIA 最先进的复合 AI 合成数据生成工具 NeMo Data Designer 构建，反映了官方的地区人口统计和地理统计数据。其目的并非重塑真实的人，而是帮助开发者测试其系统是否能反映他们所声称服务的用户、语言、地区和职业。基于 Nemotron-Personas-USA 构建的衍生数据集 Privasis 展示了这一方向的潜力，它在医疗、金融、法律和社会语境中层叠了保护隐私的合成记录。
Last month at VivaTech in Paris, we launched our tenth country in the collection, which now represents more than 2.4B people.上个月在巴黎的 VivaTech 大会上，我们发布了该系列的第十个国家数据，目前已覆盖超过 24 亿人口。
When quality is local, only people who know that locality can build it — regional researchers, native speakers, subject-matter experts, stakeholders who can inspect and correct alongside you. That's learning in public: not releasing data in isolation, but building it collaboratively.当质量具有地域性时，只有了解当地情况的人才能构建出高质量的数据——这包括区域研究人员、母语使用者、领域专家以及能与您共同审查和纠正问题的利益相关者。这就是“在公众视野中学习”：不是孤立地发布数据，而是协同构建。
Ground Truths 基本事实
Synthetic data needs to be integrated as part of a system of data sources. There are tradeoffs. It can reduce risk, but it does not remove the need for grounding, lineage, curation, evaluation, and human judgment.合成数据需要作为数据源系统的一部分进行整合。这其中存在权衡：它能降低风险，但不能取代对溯源、筛选、评估和人类判断的需求。
One useful way to think about this is with "synthetic thresholds": points where data can no longer be treated as purely real. That line is not always obvious. Real workflows, human feedback, model-generated traces, simulated users, and synthetic labels can all become intertwined. The answer is not to pretend synthetic data is fake or harmless. It is to document what was generated, what was grounded, what was reviewed, and what the data is meant to test. As more AI systems are trained on artificial information, we need better shared habits for inspecting it, documenting it, and debating these technologies in public.思考这一问题的有效方式是引入“合成阈值”：即数据不再能被视为纯粹真实数据的临界点。这条界限并不总是显而易见的。真实工作流、人类反馈、模型生成的轨迹、模拟用户和合成标签可能会交织在一起。解决之道不是假装合成数据是虚假的或无害的，而是记录下哪些是生成的、哪些是基于事实的、哪些经过了审查，以及这些数据的测试目的。随着越来越多的 AI 系统在人工信息上进行训练，我们需要建立更好的共享习惯来对其进行审查、记录，并在公共领域讨论这些技术。
Quality also means different things in different contexts. Reasoning data needs harder problems and cleaner traces. Persona data needs distributional fidelity and local review. Agentic workflows need task diversity, failure coverage, and recovery paths. The field is still more craft than formula.质量在不同语境下也有不同的含义。推理数据需要更难的问题和更清晰的轨迹；角色数据需要分布保真度和本地审查；智能体工作流则需要任务多样性、故障覆盖率和恢复路径。这一领域目前仍是“手艺”多于“公式”。
That is why open methods matter. Synthetic data is not just about generating more examples. It is about asking better questions, and making it possible for parties who otherwise could not sit at the same table: companies without giving away their secrets, governments without compromising privacy, and researchers without waiting for permission that may never come.这就是开放方法的重要性所在。合成数据不仅仅是为了生成更多示例，更是为了提出更好的问题，并使原本无法坐在一起的各方能够合作：公司无需泄露秘密，政府无需损害隐私，研究人员也无需等待那可能永远不会到来的许可。
The scarce resource in AI is not tokens. It is trust between organizations. Synthetic data is one of the few tools we have for building it.AI 领域稀缺的资源不是 Token，而是机构之间的信任。合成数据是我们构建这种信任的少数工具之一。
We hosted a livestream on Tuesday, July 7, 2026 on Why Open Data Matters with an amazing panel. It is worth checking out, along with the Nemotron data collections on Hugging Face.我们于 2026 年 7 月 7 日星期二举办了一场关于“为何开放数据至关重要”的直播，并邀请了出色的嘉宾小组。欢迎观看直播，并查看 Hugging Face 上的 Nemotron 数据集。
Stay up to date on NVIDIA Nemotron by subscribing to NVIDIA news and following NVIDIA AI on LinkedIn, X, YouTube, and the Nemotron channel on Discord.如需了解 NVIDIA Nemotron 的最新动态，请订阅 NVIDIA 新闻，并关注 LinkedIn、X、YouTube 上的 NVIDIA AI 以及 Discord 上的 Nemotron 频道。
Access open Nemotron Models on Hugging Face and a collection of NIM microservices and Developer Examples on build.nvidia.com.访问 Hugging Face 获取开放的 Nemotron 模型，并在 build.nvidia.com 探索 NIM 微服务和开发者示例。
