构建开源终端任务数据集，显著提升大模型终端操作能力。
On Data Engineering for Scaling LLM Terminal Capabilities
- 设计轻量级合成任务生成管道，支持基于种子和技能的任务构建。
- 32B模型在终端基准上准确率从3.4%提升至27.4%，接近更大模型性能。
- 开源模型与数据集，推动终端智能体研究发展。
尽管大语言模型在终端能力方面取得快速进展，但其训练数据策略仍大多未公开。本文通过系统研究终端智能体的数据工程实践，提出两项关键贡献：(1) Terminal-Task-Gen，一个轻量级的合成任务生成流水线，支持基于种子和技能的任务构造；(2) 对数据与训练策略的全面分析，包括过滤、课程学习、长上下文训练及扩展行为。该流程生成了Terminal-Corpus，一个大规模开源终端任务数据集。基于此数据集，我们训练了Nemotron-Terminal系列模型，以Qwen3（8B、14B、32B）为基础，实现显著提升：在Terminal-Bench 2.0上，Nemotron-Terminal-8B准确率从2.5%升至13.0%，14B从4.0%升至20.2%，32B从3.4%升至27.4%，性能媲美更大模型。为加速该领域研究，我们已在Hugging Face开源模型检查点及大部分合成数据集。
原文摘要 · Abstract (English)
Despite rapid recent progress in the terminal capabilities of large language models, the training data strategies behind state-of-the-art terminal agents remain largely undisclosed. We address this gap through a systematic study of data engineering practices for terminal agents, making two key contributions: (1) Terminal-Task-Gen, a lightweight synthetic task generation pipeline that supports seed-based and skill-based task construction, and (2) a comprehensive analysis of data and training strategies, including filtering, curriculum learning, long context training, and scaling behavior. Our pipeline yields Terminal-Corpus, a large-scale open-source dataset for terminal tasks. Using this dataset, we train Nemotron-Terminal, a family of models initialized from Qwen3(8B, 14B, 32B) that achieve substantial gains on Terminal-Bench 2.0: Nemotron-Terminal-8B improves from 2.5% to 13.0% Nemotron-Terminal-14B improves from 4.0% to 20.2%, and Nemotron-Terminal-32B improves from 3.4% to 27.4%, matching the performance of significantly larger models. To accelerate research in this domain, we open-source our model checkpoints and most of our synthetic datasets at https://huggingface.co/collections/nvidia/nemotron-terminal.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。