id
stringlengths 64
64
| text
stringlengths 131k
1.38M
|
|---|---|
00e65818883eddac9ef44659bd14fd17a930c4cab1f0b47409aacc4ca45ab868
| "A.\nHow we (do not) trust: defensive profiles of epistemic trust and its disruptions\nOpen access\n(...TRUNCATED)
|
220dfafb8f4c4d31ec2b0239a646f0fc7ea2ae6f247295427c8a6dc9ca805a16
| "中华人民共和国民法典\n\n（2020年5月28日第十三届全国人民代表大会第三次(...TRUNCATED)
|
104c4fca5d2cbafec3288b72ad10271a3807ffc29df957b62bb9413a924229ac
| "\n\n\n\n\n\n创意信息技术股份有限公司\n\n\n2022 年年度报告\n\n\n\n\n\n【2023 年 4(...TRUNCATED)
|
6db1aa1500ed165b075286a7b85157515434ab26590cfbecf3892b64d92e0805
| "The Red-Pilling of Kitson The boutique that defined early-aughts L.A. style has taken an … unexpe(...TRUNCATED)
|
e579ca3ffe4a238534e3c9396dfe99b847f348cbb10a87af9b511e4c23dce71b
| "Adventure\nHe was a very sick white man. He rode pick-a-back on a woolly- headed, black-skinned sav(...TRUNCATED)
|
851f08011a7d9a6f2a4da3b1500d1ab47b657fdb1eed1449670c523a62c9742f
| "UNITED STATES\nSECURITIES AND EXCHANGE COMMISSION\nWashington, D.C. 20549\n________________________(...TRUNCATED)
|
df9d087642541e4e420c5d3251b682db523c3873c7fb534f72744b478047b94a
| "INTRODUCTION\nThe duel between Nietzsche and civilisation is long since over; and that high poet an(...TRUNCATED)
|
07b656dcf9a776d8cbd0daa3e93970e5458ac4c9bc179dfa7d6aeaaeadcd1e83
| "1\tOne of the other reviewers has mentioned that after watching just 1 Oz episode you'll be hooked.(...TRUNCATED)
|
86569f634539c9e13a7d1a02082374893ed19bd489e761190fff83f9df2bcebb
| "第1段\n第一回 秋水望穿采菱舟去 栏杆依遍拂面香来\n 话说一年之中，(...TRUNCATED)
|
b655d1e22a3852761aeec5b3bf48d493955f901ed7bd648549bcce0ab06fb4f5
| "The Project Gutenberg eBook of The Confessions of Jean Jacques Noah — Complete\n\nThis ebook is (...TRUNCATED)
|
End of preview. Expand
in Data Studio
LongBench Pro 128k+
This dataset is a filtered subset of LongBench Pro (caskcsg/LongBench-Pro).
What is included
Only examples whose token_length field is one of:
128k256k
Columns
This repo keeps only:
id: example identifier (copied from source)text: the originalcontextfield (renamed fromcontext→text)
All other fields from the source dataset are dropped.
Intended use
Use this dataset when you want to benchmark long-context behavior specifically at ≥128k length buckets, while keeping the input surface minimal (id, text).
Provenance / attribution
Source dataset: caskcsg/LongBench-Pro.
This repo contains a derived subset. Please consult the source dataset card for:
- full task definitions
- original annotations/fields
- licensing/usage terms
Reproducibility
The filtering logic and transformation used to build this dataset are contained in create_dataset.py in this repo.
- Downloads last month
- 17