On Data Engineering for Scaling LLM Terminal Capabilities
Paper • 2602.21193 • Published • 102
conversations list | agent string | model string | model_provider string | date string | task string | episode string | run_id string | trial_name string | enable_thinking bool | source string | difficulty string |
|---|---|---|---|---|---|---|---|---|---|---|---|
[
{
"content": "You are an AI assistant tasked with solving command-line tasks in a Linux environment. You will be given a task description and the output from previously executed commands. Your goal is to solve the task by providing batches of shell commands.\n\nFormat your response as JSON with the following st... | terminus-2 | deepseek-ai/DeepSeek-V3.2 | hosted_vllm | 2025-12-31T15:49:37.661438 | software_engineering_task_1210 | episode-6 | 8dd2b9cb-393d-4c79-9e0c-fa185b64f444 | software_engineering_task_1210__5NPxtBh | true | software_engineering | easy |
[
{
"content": "You are an AI assistant tasked with solving command-line tasks in a Linux environment. You will be given a task description and the output from previously executed commands. Your goal is to solve the task by providing batches of shell commands.\n\nFormat your response as JSON with the following st... | terminus-2 | deepseek-ai/DeepSeek-V3.2 | hosted_vllm | 2025-12-31T15:35:19.856411 | software_engineering_task_3534 | episode-6 | 8dd2b9cb-393d-4c79-9e0c-fa185b64f444 | software_engineering_task_3534__SWCjTJR | true | software_engineering | easy |
[
{
"content": "You are an AI assistant tasked with solving command-line tasks in a Linux environment. You will be given a task description and the output from previously executed commands. Your goal is to solve the task by providing batches of shell commands.\n\nFormat your response as JSON with the following st... | terminus-2 | deepseek-ai/DeepSeek-V3.2 | hosted_vllm | 2025-12-31T10:27:02.180804 | software_engineering_task_1631 | episode-7 | 8dd2b9cb-393d-4c79-9e0c-fa185b64f444 | software_engineering_task_1631__bHnbcoL | true | software_engineering | easy |
[{"content":"You are an AI assistant tasked with solving command-line tasks in a Linux environment. (...TRUNCATED) | terminus-2 | deepseek-ai/DeepSeek-V3.2 | hosted_vllm | 2025-12-31T13:41:23.593607 | software_engineering_task_3521 | episode-12 | 8dd2b9cb-393d-4c79-9e0c-fa185b64f444 | software_engineering_task_3521__NsWtvKU | true | software_engineering | easy |
[{"content":"You are an AI assistant tasked with solving command-line tasks in a Linux environment. (...TRUNCATED) | terminus-2 | deepseek-ai/DeepSeek-V3.2 | hosted_vllm | 2025-12-31T14:01:20.557852 | software_engineering_task_0906 | episode-9 | 8dd2b9cb-393d-4c79-9e0c-fa185b64f444 | software_engineering_task_0906__SdEZvAy | true | software_engineering | easy |
[{"content":"You are an AI assistant tasked with solving command-line tasks in a Linux environment. (...TRUNCATED) | terminus-2 | deepseek-ai/DeepSeek-V3.2 | hosted_vllm | 2025-12-31T13:25:54.979232 | software_engineering_task_1800 | episode-10 | 8dd2b9cb-393d-4c79-9e0c-fa185b64f444 | software_engineering_task_1800__q9Wpoac | true | software_engineering | easy |
[{"content":"You are an AI assistant tasked with solving command-line tasks in a Linux environment. (...TRUNCATED) | terminus-2 | deepseek-ai/DeepSeek-V3.2 | hosted_vllm | 2025-12-31T12:39:19.245458 | software_engineering_task_3970 | episode-24 | 8dd2b9cb-393d-4c79-9e0c-fa185b64f444 | software_engineering_task_3970__MfhyTTW | true | software_engineering | easy |
[{"content":"You are an AI assistant tasked with solving command-line tasks in a Linux environment. (...TRUNCATED) | terminus-2 | deepseek-ai/DeepSeek-V3.2 | hosted_vllm | 2025-12-31T12:05:29.518024 | software_engineering_task_1266 | episode-6 | 8dd2b9cb-393d-4c79-9e0c-fa185b64f444 | software_engineering_task_1266__4L3fCHJ | true | software_engineering | easy |
[{"content":"You are an AI assistant tasked with solving command-line tasks in a Linux environment. (...TRUNCATED) | terminus-2 | deepseek-ai/DeepSeek-V3.2 | hosted_vllm | 2025-12-31T11:30:16.351771 | software_engineering_task_4879 | episode-5 | 8dd2b9cb-393d-4c79-9e0c-fa185b64f444 | software_engineering_task_4879__qk9y4vC | true | software_engineering | easy |
[{"content":"You are an AI assistant tasked with solving command-line tasks in a Linux environment. (...TRUNCATED) | terminus-2 | deepseek-ai/DeepSeek-V3.2 | hosted_vllm | 2025-12-31T07:57:04.577445 | software_engineering_task_1127 | episode-10 | 8dd2b9cb-393d-4c79-9e0c-fa185b64f444 | software_engineering_task_1127__zoc7vpo | true | software_engineering | easy |
Per-source partition of nvidia/Nemotron-Terminal-Corpus,
filtered to source == "software_engineering". The difficulty column preserves the original
easy / medium / mixed split (na for the dataset_adapters/* files, which
did not carry a difficulty label).
Partitioning scheme:
dataset_adapters/{code,math,swe}.parquetdebugging, security, …) — rows from
synthetic_tasks/skill_based/{easy,medium,mixed}/{skill}/data_filtered.parquetSame as the source dataset (conversations, agent, model, model_provider,
date, task, episode, run_id, trial_name, enable_thinking) plus:
source — the partition key ("software_engineering" throughout this repo)difficulty — easy / medium / mixed / naoriginal_source — only present in adapters_code; preserves the original
source column value (OpenCodeReasoning or synthetic) from the upstream file.@misc{pi2026dataengineeringscalingllm,
title={On Data Engineering for Scaling LLM Terminal Capabilities},
author={Renjie Pi and Grace Lam and Mohammad Shoeybi and Pooya Jannaty and Bryan Catanzaro and Wei Ping},
year={2026},
eprint={2602.21193},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2602.21193},
}
Original dataset license: CC-BY-4.0.