Datasets:
segments listlengths 2 2 |
|---|
[
{
"label": false,
"text": "<bos><|turn>system\n<|think|>\nYou are a capable assistant that solves tasks by using the tools you have been given. You help the user finish the request completely, with evidence from tools when the request needs facts, current data, calculations, lookups, or side effects.\n\nThe... |
[
{
"label": false,
"text": "<bos><|turn>system\n<|think|>\nYou are a capable assistant that solves tasks by using the tools you have been given. You help the user finish the request completely, with evidence from tools when the request needs facts, current data, calculations, lookups, or side effects.\n\nThe... |
[
{
"label": false,
"text": "<bos><|turn>system\n<|think|>\n<turn|>\n<|turn>user\n당신은 도움이 되고 해가 없는 조수입니다. 아래 지시에 응답하기 전에 단계별로 생각해야 합니다.\n\n파이썬 프로그래밍 언어만 사용해 주세요.\n\n최종 솔루션 코드 블록에는 다음 형식으로 ```python만 사용해야 합니다:\n```python\n# 여기에 코드 작성\n```\n\n가격 차이\n\n문제 설명\n\n<!-- 여기에 어떤 이미지를 넣어야 할지 모르겠네요... 물, 탄산음료, 레모네이드 이미지... |
[
{
"label": false,
"text": "<bos><|turn>system\n<|think|>\nYou are a capable assistant that solves tasks by using the tools you have been given. You help the user finish the request completely, with evidence from tools when the request needs facts, current data, calculations, lookups, or side effects.\n\nThe... |
[
{
"label": false,
"text": "<bos><|turn>system\n<|think|>\nYou are a capable assistant that solves tasks by using the tools you have been given. You help the user finish the request completely, with evidence from tools when the request needs facts, current data, calculations, lookups, or side effects.\n\nThe... |
[
{
"label": false,
"text": "<bos><|turn>system\n<|think|>\n<turn|>\n<|turn>user\nResponda à seguinte questão de múltipla escolha. A última linha da sua resposta deve estar no seguinte formato: 'Resposta: \\boxed{A/B/C/D}' (exemplo: 'Resposta: \\boxed{A}').\n\nNo eletrocatalisador de calcogeneto de metal em c... |
[
{
"label": false,
"text": "<bos><|turn>system\n<|think|>\n<turn|>\n<|turn>user\nनिम्नलिखित बहुविकल्पीय प्रश्न का उत्तर दें। आपकी प्रतिक्रिया की अंतिम पंक्ति निम्न प्रारूप में होनी चाहिए: 'उत्तर: A/B/C/D/E/F/G/H/I/J' (जैसे 'उत्तर: A')।\n\nउच्च-ऊर्जा भौतिकी प्रयोगों में रैखिक त्वरकों की तुलना में वृत्ताकार कण... |
[
{
"label": false,
"text": "<bos><|turn>system\n<|think|>\n<turn|>\n<|turn>user\n당신은 도움이 되고 해롭지 않은 어시스턴트입니다. 아래 지시사항에 응답하기 전에 단계별로 생각해야 합니다.\n\nC++ 프로그래밍 언어만 사용하세요.\n\n최종 솔루션 코드 블록에는 다음 형식으로 ```cpp를 사용해야 합니다:\n```cpp\n// 여기에 코드 작성\n```\n\n회문 부분 문자열\n\n문제 설명\n\n문자열 ${s}가 주어집니다. ${s}의 부분 문자열 중 회문으로 재배열할 수 있는 부... |
[
{
"label": false,
"text": "<bos><|turn>system\n<|think|>\n<turn|>\n<|turn>user\nआप एक सहायक और हानिरहित सहायक हैं। नीचे दिए गए निर्देश का जवाब देने से पहले आपको चरण-दर-चरण सोचना चाहिए।\n\nकृपया केवल C++ प्रोग्रामिंग भाषा का उपयोग करें।\n\nअंतिम समाधान कोड ब्लॉक के लिए आपको निम्नलिखित प्रारूप में ```cpp का उ... |
[
{
"label": false,
"text": "<bos><|turn>system\n<|think|>\nYou are a capable assistant that solves tasks by using the tools you have been given. You help the user finish the request completely, with evidence from tools when the request needs facts, current data, calculations, lookups, or side effects.\n\nThe... |
[
{
"label": false,
"text": "<bos><|turn>system\n<|think|>\nYou are the organization's front-line agent for whatever domain the current policy describes. That domain may be retail, travel, education, healthcare, construction, aviation, municipal services, marketplaces, or any other vertical in the policy. The... |
[
{
"label": false,
"text": "<bos><|turn>system\n<|think|>\nYou are a capable assistant that solves tasks by using the tools you have been given. You help the user finish the request completely, with evidence from tools when the request needs facts, current data, calculations, lookups, or side effects.\n\nThe... |
End of preview. Expand in Data Studio
Nemotron-SFT-Agentic-Multilingual-segments
First 1,000,000 rows of a merged, interleaved mix of:
Nemotron-SFT-Agentic-v2-segments-filtered-1m-en-final.jsonl(784,987 rows)Nemotron-SFT-Multilingual-v2-segments.jsonl(370,081 rows)
The two sources were split into 1,000-line chunks, chunk order shuffled (seed 42), then round-robin merged with paste -d '\n'. This Hub snapshot keeps only the first 1,000,000 lines of that shuffle.
Each row is:
{"segments": [{"label": false, "text": "..."}, {"label": true, "text": "..."}]}
Load
from datasets import load_dataset
ds = load_dataset("AlexHung29629/Nemotron-SFT-Agentic-Multilingual-segments", split="train")
- Downloads last month
- 93