Dataset Viewer
Auto-converted to Parquet Duplicate
text
stringlengths
310
8.3k
writing_style_average
float64
-3.5
3.37
facts_and_trivia_average
float64
-3.77
3.96
educational_value_average
float64
-3.86
3.99
required_expertise_average
float64
-3.75
4.36
writing_style_chunks
sequencelengths
2
2
facts_and_trivia_chunks
sequencelengths
2
2
educational_value_chunks
sequencelengths
2
2
required_expertise_chunks
sequencelengths
2
2
length
int64
1.02k
1.02k
chunk_lengths
sequencelengths
2
2
input_ids
sequencelengths
1.02k
1.02k
document_index
int64
75
1.18M
document_position
int64
0
1.47M
source_domain
stringclasses
7 values
cluster_id
int64
-1
24
cluster_no
int64
-1
25
Louisville baseball's season comes to an end on Vanderbilt rally in College World Series Vanderbilt rallied in the top of the ninth inning to end Louisville's historic postseason run. Louisville baseball's season comes to an end on Vanderbilt rally in College World Series Vanderbilt rallied in the top of the ninth inni...
-0.3776
-0.390198
-0.992246
-1.453781
[ -0.9665603514126393, 0.21136051752629473 ]
[ -0.6484963602830048, -0.13190001457515102 ]
[ -1.5074203559848984, -0.477072439422539 ]
[ -1.7594416817878897, -1.148119668748462 ]
1,024
[ 512, 512 ]
[ 5899, 4909, 21573, 29915, 29879, 4259, 5304, 304, 385, 1095, 373, 478, 3825, 29890, 2782, 364, 635, 297, 6346, 2787, 10488, 13, 29963, 3825, 29890, 2782, 364, 284, 2957, 297, 278, 2246, 310, 278, 17081, 386, 297, 1076, 304, 1095, 5899, ...
75
0
CommonCrawl
5
7
he said looking at McDonnell. "But, yeah, I mean, that's something, from day one, that was our goal, to be the best team he's ever coached. To accomplish that is something super special, and it takes a special group." This year's team has been built on its pitching depth. It's one of the main reason why many thought th...
0.153834
-0.029792
-0.196818
-0.874976
[ 0.11411356206738273, 0.19355473694931086 ]
[ 0.2020522158391115, -0.2616366510053567 ]
[ -0.046545525363559326, -0.3470900868715952 ]
[ -0.8151442744120928, -0.9348073067517255 ]
1,024
[ 512, 512 ]
[ 540, 1497, 3063, 472, 4052, 29928, 3409, 514, 29889, 376, 6246, 29892, 21915, 29892, 306, 2099, 29892, 393, 29915, 29879, 1554, 29892, 515, 2462, 697, 29892, 393, 471, 1749, 7306, 29892, 304, 367, 278, 1900, 3815, 540, 29915, 29879, 3926,...
75
1,024
CommonCrawl
5
7
Home / Press Release / McKinsey & Company's new research shows the metaverse could grow up to $5 trillion in value by 2030 McKinsey & Company's new research shows the metaverse could grow up to $5 trillion in value by 2030 in Press Release June 15, 2022 256 Views | Short URL 'Value Creation in the Metaverse' report exa...
1.038644
1.885243
1.155633
0.753482
[ 0.7537519874098038, 1.323536965873288 ]
[ 1.6918551954887975, 2.078630969051696 ]
[ 1.003458259145528, 1.3078071821916404 ]
[ 0.5245614137381425, 0.9824025809506501 ]
1,024
[ 512, 512 ]
[ 8778, 847, 5254, 23708, 847, 24053, 262, 7759, 669, 6938, 29915, 29879, 716, 5925, 3697, 278, 12700, 3901, 1033, 6548, 701, 304, 395, 29945, 534, 453, 291, 297, 995, 491, 29871, 29906, 29900, 29941, 29900, 13, 27297, 29968, 262, 7759, 6...
77
0
CommonCrawl
15
4
Digital / ICT Healthcare and Biomedical Urban Solutions Infrastructure Hub Quality & Standards E-Alert for Standards For Conformity Assessment Bodies Mutual Recognition Arrangements (MRAs) For Local companies Loans & Insurance SME Venture Loan SME Equipment and Factory Loan Bridging Loan for Marine and Offshore Enginee...
0.622092
0.408198
0.311064
-0.405103
[ -0.19988453003058307, 1.444068403625179 ]
[ -0.2738073296439462, 1.0902039921687319 ]
[ 0.20200609512409906, 0.4201228233071462 ]
[ -0.5062315550570629, -0.3039750167004011 ]
1,024
[ 512, 512 ]
[ 15918, 847, 306, 1783, 13, 3868, 4298, 18020, 322, 3457, 27067, 936, 13, 29965, 29878, 2571, 4956, 17925, 13, 797, 14867, 12425, 14533, 13, 24399, 537, 669, 6679, 3163, 13, 29923, 29899, 16649, 363, 6679, 3163, 13, 2831, 1281, 689, 537,...
79
0
CommonCrawl
15
4
consumer needs. Ms Tan, the one who takes care of business development, finance and logistics, modestly attributes Love, Bonito's success to first-mover advantage and "being blessed", but a look at the company's progress over nine years uncovers several smart moves. Love, Bonito churns out approximately 70 new in-house...
1.471462
0.825191
0.544399
-0.279912
[ 1.6522590688329906, 1.2906647555773179 ]
[ 0.7535658188825048, 0.8968161053872823 ]
[ 0.653456997642499, 0.4353402694594518 ]
[ -0.47826682467334436, -0.0815578587647795 ]
1,024
[ 512, 512 ]
[ 21691, 4225, 29889, 13, 29924, 29879, 15066, 29892, 278, 697, 1058, 4893, 2562, 310, 5381, 5849, 29892, 1436, 749, 322, 1480, 6765, 29892, 878, 342, 368, 8393, 8155, 29892, 8396, 2049, 29915, 29879, 2551, 304, 937, 29899, 29885, 957, 1063...
79
1,024
CommonCrawl
15
4
BATON ROUGE – The LSU Office of the Dean of Students is proud to present the Tiger Twelve Class of 2014 at a ceremony to be held Saturday, April 26, at 2 p.m. in the Rotunda of the LSU Business Education Complex. • De André Beadle, a native of Houston, majoring in sociology and minoring in political science. While at L...
0.394897
0.775053
0.932444
-0.339744
[ 0.3510674112841683, 0.438726638740089 ]
[ 0.7034282186058327, 0.8466785051106102 ]
[ 0.8817186899270831, 0.9831683309424539 ]
[ -0.49907778681936743, -0.1804099289583891 ]
1,024
[ 512, 512 ]
[ 350, 1299, 1164, 16641, 29965, 1692, 785, 450, 365, 14605, 11367, 310, 278, 23263, 310, 5088, 1237, 338, 22314, 304, 2198, 278, 323, 4087, 8168, 13841, 4134, 310, 29871, 29906, 29900, 29896, 29946, 472, 263, 25672, 304, 367, 4934, 24211, ...
81
0
C4
21
17
in Louisiana and around the world to increase access to healthcare and decrease rates of HIV/AIDS. • Nathan Irby, a native of Shreveport majoring in marketing and minoring in entrepreneurship. Irby is currently working with the Office of the Baton Rouge Mayor-President and local employers to place 100 at-risk youth in ...
0.323674
0.837725
0.937516
-0.255199
[ 0.23875402610626995, 0.4085937793021162 ]
[ 0.5494341606131969, 1.1260165637949262 ]
[ 0.71939926430249, 1.1556327206685844 ]
[ -0.40542845716226356, -0.10497019117905546 ]
1,024
[ 512, 512 ]
[ 297, 28838, 322, 2820, 278, 3186, 304, 7910, 2130, 304, 9045, 18020, 322, 23806, 19257, 310, 379, 5667, 29914, 29909, 1367, 29903, 29889, 13, 30119, 27650, 6600, 1609, 29892, 263, 7531, 310, 1383, 276, 345, 637, 4655, 292, 297, 9999, 29...
81
1,024
C4
21
17
UK Government Publishes List Of Ubuntu 18.04 LTS Security Suggestions Your Blog » UK Government Publishes List Of Ubuntu 18.04 LTS Security Suggestions Previous: The Man Accountable For Passwords Says Tips Was Wrong Next: Facebook Makes Users Getting Ads To Offer Social Safety Quantity This will be much more helpful fo...
-0.847399
-0.979763
-0.608957
-1.036261
[ -1.1418788063244807, -0.5529183718550136 ]
[ -0.9690188763374444, -0.9905064193131611 ]
[ -0.8200990381057615, -0.3978149073792806 ]
[ -1.3172087361848994, -0.7553127582422765 ]
1,024
[ 512, 512 ]
[ 10261, 10354, 12904, 267, 2391, 4587, 8294, 29871, 29896, 29947, 29889, 29900, 29946, 365, 9375, 14223, 317, 12981, 2297, 13, 10858, 350, 1188, 2047, 10261, 10354, 12904, 267, 2391, 4587, 8294, 29871, 29896, 29947, 29889, 29900, 29946, 365, ...
87
0
CommonCrawl
12
5
In late September 2014, thousands of people started to occupy major streets in Hong Kong which lasted for more than two months. The protesters followed the call of the Occupy Central with Love and Peace movement, which had threatened civil disobedience if real democracy was denied. This had happened on August 31st when...
1.345452
1.530699
1.510706
1.044835
[ 1.3673665796012484, 1.323536965873288 ]
[ 2.0070058257993075, 1.0543914205425375 ]
[ 1.5817212129331415, 1.4396917155116224 ]
[ 1.211323164556904, 0.8783477702205348 ]
1,024
[ 512, 512 ]
[ 512, 5683, 3839, 29871, 29906, 29900, 29896, 29946, 29892, 17202, 310, 2305, 4687, 304, 6919, 29891, 4655, 19756, 297, 12798, 12026, 607, 1833, 287, 363, 901, 1135, 1023, 7378, 29889, 450, 10021, 414, 5643, 278, 1246, 310, 278, 16117, 786...
88
0
C4
6
2
were sold at production costs. In addition, you could get free water, snacks, or hot soup from supply stations. Many Hongkongers who did not dare to participate, donated large amounts of resources for the supply stations that supported it. Another important difference was the fact that the Umbrella Movement was a site ...
1.52077
1.552186
1.556359
0.558379
[ 1.6632164722649807, 1.3783239830332386 ]
[ 1.8494305106440525, 1.254941821649226 ]
[ 2.088969418009995, 1.0237481873486023 ]
[ 0.6546299271507867, 0.4621285273000733 ]
1,024
[ 512, 512 ]
[ 892, 5239, 472, 5802, 21544, 29889, 512, 6124, 29892, 366, 1033, 679, 3889, 4094, 29892, 5807, 26514, 29892, 470, 7375, 22300, 515, 11421, 16355, 29889, 9267, 12798, 29895, 549, 414, 1058, 1258, 451, 23222, 304, 5221, 403, 29892, 1016, 63...
88
1,024
C4
6
2
_About & FAQs _Britt on CinemaBlend _Movie Reviews _Movies on Netflix _Must-See Movies _Television Reviews _TV Movie Reviews _Game of Thrones _Best TV on Netflix _Playlists _Music Reviews _Netflix _Amazon Prime Video _Acorn TV Interview with Sam Batt (Founder of Battfest) Eclectic Pop was fortunate to have Battfest fou...
0.193212
-0.40139
-0.141021
-1.37704
[ 0.010703067177976314, 0.375721569006146 ]
[ -0.8400936184831447, 0.037314386358617384 ]
[ -0.4282497996838918, 0.14620879256564515 ]
[ -1.6033594656927168, -1.1507210390167149 ]
1,024
[ 512, 512 ]
[ 903, 28173, 669, 13515, 29984, 29879, 13, 29918, 29933, 12123, 373, 27306, 10358, 355, 13, 29918, 18749, 830, 7406, 13, 29918, 29924, 586, 583, 373, 12670, 20157, 29916, 13, 29918, 29924, 504, 29899, 13393, 14104, 583, 13, 29918, 29911, 6...
92
0
CommonCrawl
20
25
internet enables us but it also makes it difficult to stand out from the crowd, because there are so many great indie artists out there In the current climate of how people discover new music, what do you see as the best way for rising artists to be discovered? In my opinion sharing reciprocal promotions is the way for...
0.341137
-0.648496
-0.697091
-1.612464
[ 0.7318371805458236, -0.04956265169796915 ]
[ -0.08802961433306292, -1.2089631062329467 ]
[ 0.001643054118741791, -1.3958257508679905 ]
[ -1.0544703390913581, -2.1704581841718453 ]
1,024
[ 512, 512 ]
[ 8986, 28936, 502, 541, 372, 884, 3732, 372, 5189, 304, 2317, 714, 515, 278, 19174, 29892, 1363, 727, 526, 577, 1784, 2107, 1399, 347, 17906, 714, 727, 13, 797, 278, 1857, 23622, 310, 920, 2305, 6523, 716, 4696, 29892, 825, 437, 366, ...
92
1,024
CommonCrawl
20
25
#include <Xm/Xm.h> #include <Xbae/MatrixP.h> #include <Xbae/Actions.h> #include <Xbae/Utils.h> #include <Xbae/Clip.h> /* * Public.c created by Andrew Lister (7 August, 1995) */ /* * Public interface to set_cell method */ void XbaeMatrixSetCell(w, row, column, value ) Widget w; int row; int column; const String ...
-1.569218
-1.645876
-0.417629
0.345067
[ -1.5363453298761236, -1.6020897504680642 ]
[ -1.6029013941210848, -1.6888515660239514 ]
[ -0.5262121092893592, -0.3090464714908312 ]
[ 0.44131756515405024, 0.24881616530333678 ]
1,024
[ 512, 512 ]
[ 29871, 13, 13, 29937, 2856, 529, 29990, 29885, 29914, 29990, 29885, 29889, 29882, 29958, 13, 29937, 2856, 529, 29990, 2291, 29872, 29914, 14609, 29925, 29889, 29882, 29958, 13, 29937, 2856, 529, 29990, 2291, 29872, 29914, 26525, 29889, 29882,...
93
0
Github
-1
-1
, NULL); /* * Call the deselect_cell method */ (*((XbaeMatrixWidgetClass) XtClass(w))->matrix_class.deselect_cell) ((XbaeMatrixWidget) w, row, column); } /* * Public interface to deselect_row method */ void XbaeMatrixDeselectRow(w, row) Widget w; int row; { /* * Make sure w is a Matrix o...
-1.566478
-1.452489
-0.270369
0.220201
[ -1.711663784787965, -1.4212925938402277 ]
[ -1.6673640230482347, -1.2376131635339023 ]
[ -0.3749887381508222, -0.16574885355661997 ]
[ 0.04590928437961182, 0.39449290032549833 ]
1,024
[ 512, 512 ]
[ 29892, 4265, 416, 13, 13, 1678, 4949, 13, 268, 334, 8251, 278, 553, 15436, 29918, 3729, 1158, 13, 268, 3776, 13, 1678, 3070, 3552, 29990, 2291, 29872, 14609, 8801, 2385, 29897, 1060, 29873, 2385, 29898, 29893, 876, 976, 5344, 29918, 199...
93
1,024
Github
-1
-1
) Widget w; int position; int num_rows; { /* * Make sure w is a Matrix or a subclass */ XtCheckSubclass(w, xbaeMatrixWidgetClass, NULL); /* * Call the delete_rows method */ (*((XbaeMatrixWidgetClass) XtClass(w))->matrix_class.delete_rows) ((XbaeMatrixWidget) w, position, num_rows);...
-1.426771
-1.255519
-0.071908
0.436115
[ -1.4979944178641584, -1.3555481732482872 ]
[ -1.3307258497620076, -1.1803130489319913 ]
[ 0.039686669499505826, -0.18350254073430985 ]
[ 0.4985477110556137, 0.37368193817947526 ]
1,024
[ 512, 512 ]
[ 29897, 13, 8801, 281, 29936, 13, 524, 2602, 29936, 13, 524, 954, 29918, 5727, 29936, 13, 29912, 13, 1678, 4949, 13, 268, 334, 8561, 1854, 281, 338, 263, 22513, 470, 263, 19481, 13, 268, 3776, 13, 1678, 1060, 29873, 5596, 4035, 1990, ...
93
2,048
Github
-1
-1
, num_colors) Widget w; int position; Pixel *colors; int num_colors; { /* * Make sure w is a Matrix or a subclass */ XtCheckSubclass(w, xbaeMatrixWidgetClass, NULL); /* * Call the set_row_colors method */ (*((XbaeMatrixWidgetClass) XtClass(w))->matrix_class.set_row_colors) ((XbaeMatrix...
-1.317197
-1.174941
0.042223
0.527163
[ -1.322675962952317, -1.311718559520327 ]
[ -1.1480817344684164, -1.201800591907708 ]
[ 0.22736850537794176, -0.14292268432816155 ]
[ 0.4673312678365791, 0.5869943001762118 ]
1,024
[ 512, 512 ]
[ 29892, 954, 29918, 27703, 29897, 13, 8801, 281, 29936, 13, 524, 2602, 29936, 13, 29637, 334, 27703, 29936, 13, 524, 954, 29918, 27703, 29936, 13, 29912, 13, 1678, 4949, 13, 268, 334, 8561, 1854, 281, 338, 263, 22513, 470, 263, 19481, ...
93
3,072
Github
-1
-1
\section*{Introduction} The main result of the present paper is to construct a two-parameter family of Markov processes $\x{\alpha}{\theta}(t)$ in the infinite-dimensional Kingman simplex \begin{equation*} \overline{\nabla}_\infty=\biggl\{x=(x_1,x_2,\dots)\in[0,1]^\infty\colon x_1\ge x_2\ge\dots\ge0,\ {\displaystyle...
0.08107
-1.219707
0.163962
2.064573
[ 0.26614753468624514, -0.10400725000066988 ]
[ -0.9224625332233918, -1.5169512222182184 ]
[ 0.40997785920560914, -0.08205289971893909 ]
[ 2.220654828639023, 1.9084903964486768 ]
1,024
[ 512, 512 ]
[ 320, 2042, 19740, 25898, 29913, 13, 1576, 1667, 1121, 310, 278, 2198, 5650, 338, 304, 3386, 263, 1023, 29899, 15501, 13, 11922, 310, 4485, 586, 10174, 29871, 13, 4535, 29916, 741, 2312, 3331, 3416, 2119, 29873, 1262, 297, 278, 10362, 29...
99
0
ArXiv
-1
-1
}{\theta}(t)$ is ergodic with respect to the measure $\pd{\alpha}{\theta}$\/}. \smallskip Fix arbitrary $0\le\alpha<1$ and $\theta>-\alpha$. We construct the process in the simplex $\overline{\nabla}_\infty$ as a limit for a sequence of finite Markov chains with growing number of states. The state space for $n$th...
0.320935
-0.036997
0.534254
1.934504
[ 0.3455887095681733, 0.29628039412421786 ]
[ -0.609102531494191, 0.5351091319627191 ]
[ 0.5367899104748226, 0.531717428424054 ]
[ 1.8980849153756654, 1.9709232828867462 ]
1,024
[ 512, 512 ]
[ 3331, 3416, 2119, 29873, 1262, 13, 12, 275, 604, 17083, 293, 411, 3390, 304, 278, 5645, 29871, 13, 12, 4535, 15926, 741, 2312, 3331, 3416, 1042, 25558, 1836, 13, 13, 29905, 9278, 11014, 13, 29943, 861, 11472, 395, 29900, 29905, 280, 2...
99
1,024
ArXiv
-1
-1
} and other works. The results of the present paper extend some of the results of \cite{EK1} and \cite{Et} to the case $\alpha\ne0$. However, it seems that in the general case there is no connection between the model $\x{\alpha}{\theta}(t)$ and population genetics. Namely, both finite Markov chains and the process in $...
0.473653
-0.226356
0.942588
2.090586
[ 0.1113742112093852, 0.8359325131497294 ]
[ -0.30872458697948574, -0.14398675749899162 ]
[ 0.6382395514901934, 1.2469373975824178 ]
[ 1.6899752939154347, 2.491197336537323 ]
1,024
[ 512, 512 ]
[ 29913, 322, 916, 1736, 29889, 13, 1576, 2582, 310, 278, 2198, 5650, 13, 21843, 777, 310, 278, 2582, 13, 974, 320, 2036, 29912, 29923, 29968, 29896, 29913, 322, 320, 2036, 29912, 29923, 29873, 29913, 13, 517, 278, 1206, 779, 2312, 29905,...
99
2,048
ArXiv
-1
-1
}) for the operator $A$. Using this formula it is possible to prove the continuity of the sample paths of the process in $\overline{\nabla}_\infty$. In \S\ref{s5.4} we deal with the case of degenerate values of the parameters, namely, $\alpha=-\beta$, $\theta=N\beta$. In this case as a limit of up/down Markov chains ...
0.290802
-0.124738
0.952733
1.820044
[ -0.029702357977487118, 0.6113057427939328 ]
[ -0.49450230229036907, 0.24502730179054474 ]
[ 0.4860650899671372, 1.4194017873085483 ]
[ 1.52348759674725, 2.1166000179089077 ]
1,024
[ 512, 512 ]
[ 1800, 363, 278, 5455, 395, 29909, 1504, 13, 15156, 445, 7063, 372, 338, 1950, 13, 517, 6356, 278, 3133, 537, 310, 278, 4559, 10898, 310, 278, 1889, 297, 779, 6875, 741, 8511, 12806, 3411, 1504, 13, 797, 320, 29903, 29905, 999, 29912, ...
99
3,072
ArXiv
-1
-1
End of preview. Expand in Data Studio

QuRatedPajama

Paper: QuRating: Selecting High-Quality Data for Training Language Models

This dataset is a 1B token subset derived from princeton-nlp/QuRatedPajama-260B, which is a subset of cerebras/SlimPajama-627B annotated by princeton-nlp/QuRater-1.3B with sequence-level quality ratings across 4 criteria:

  • Educational Value - e.g. the text includes clear explanations, step-by-step reasoning, or questions and answers
  • Facts & Trivia - how much factual and trivia knowledge the text contains, where specific facts and obscure trivia are preferred over more common knowledge
  • Writing Style - how polished and good is the writing style in the text
  • Required Expertise: - how much required expertise and prerequisite knowledge is necessary to understand the text

This subset is useful for analysis of quality ratings. It unsupervised domain clusters for the CommonCrawl and C4 domains (a description of these clusters can be found here). We also report the quality ratings per 512 token chunk of each example.

In a pre-processing step, we split documents in into chunks of exactly 1024 tokens. We provide tokenization with the Llama-2 tokenizer in the input_ids column.

Guidance on Responsible Use:

In the paper, we document various types of bias that are present in the quality ratings (biases related to domains, topics, social roles, regions and languages - see Section 6 of the paper). Hence, be aware that data selection with QuRating could have unintended and harmful effects on the language model that is being trained. We strongly recommend a comprehensive evaluation of the language model for these and other types of bias, particularly before real-world deployment. We hope that releasing the data/models can facilitate future research aimed at uncovering and mitigating such biases. Note that the quality ratings do not measure the social or literary value of a text and should not be used for textual or demographic studies.

Citation:

@article{wettig2024qurating,
   title={QuRating: Selecting High-Quality Data for Training Language Models},
   author={Alexander Wettig, Aatmik Gupta, Saumya Malik, Danqi Chen},
   journal={arXiv preprint 2402.09739},
   year={2024}
}
Downloads last month
94

Paper for princeton-nlp/QuRatedPajama-1B_tokens_for_analysis