Datasets:

kd13
/

bookcorpus-clean

doc_id int64 0 4.09k	sent_id int64 0 385k	text stringlengths 20 1k
0	0	i wish i had a better answer to that question .
0	1	starlings , new york is not the place youd expect much to happen .
0	2	its a small quiet town , the kind where everyone knows your name .
0	3	its a place where your parents wouldnt even care if you stayed out late biking with your friends .
0	4	they dont know the half of it .
0	5	i know it all and starlings is not the place where you want to be after dark .
0	6	the only reason why no one knows this is because jason , emily , seth and i have kept it that way .
0	7	i walked along the empty road alone , occasionally waving to passing kids on bikes .
0	8	my backpack was slung over my shoulder , filled with my writing books and sketchpads .
0	9	i kept my eyes on the shadowed road , watching my every step .
0	10	usually i was more aware of my surroundings , but today i was tired and didnt care if i rammed into a tree .
0	11	i kicked a rock into the grass .
0	12	the sun was starting to set , painting the sky in brilliant oranges and reds .
0	13	it slipped down the sky , allowing the first stars to peek out from behind the bright curtain .
0	14	the remaining light cast shadows over everything , creating the illusion that there was double of everything .
0	15	the world prepared to go into its hours of unreal silence that made it seem magical , and it really was .
0	16	my phone buzzed and i awoke from my trance .
0	17	i thumbed the keypad and opened the message seth had sent me .
0	18	it read : whatcha doing ?
0	19	i flipped open the pad and wrote : walking home .
0	20	i shoved it back in my pocket and continued walking .
0	21	the sun was almost gone and shadows were starting to appear behind everything .
0	22	i looked at my own shadow , that danced behind me .
0	23	the phone buzzed again as seth sent me a reply .
0	24	do you want me to meet you ?
0	25	i glanced at the nearby woods .
0	26	my ears heard nothing and i didnt see any movement but my hands shook slightly as i typed , nothing , im fine .
0	27	i didnt open his reply because i knew what it would say .
0	28	i already knew that it was strangely quiet .
0	29	maybe i could stay home and sleep one night .
0	30	or maybe i could finish a painting tonight and watch mom do origami .
0	31	i turned the corner , and my house came into view .
0	32	i started to climb the hill when i froze .
0	33	out of my left eye i saw the outline of a creature .
0	34	it was heading through the woods in the direction of my house .
0	35	i swore under my breath and took off for the woods .
0	36	my feet crunched on the fallen branches and leaves .
0	37	i could see the creature ahead of me .
0	38	i didnt have much time .
0	39	i placed my bag on the ground and crossed my legs .
0	40	i started to grab my phone but stopped .
0	41	i didnt need to tell the others , it was just one .
0	42	and if i was lucky , it would be an easy take down .
0	43	i closed my eyes and began my mental countdown .
0	44	the wind picked up around me , as if a wind current had just appeared out of nowhere .
0	45	5 my body started to feel hot , like i was being hugged by the sun .
0	46	4 power surged through me as i was freed from my human self .
0	47	3 i felt cold , like being dipped in ice water .
0	48	the breeze picked up even harder .
0	49	2 i was hollow and could probably see my own body if i opened my eyes .
0	50	but i could never get the courage to do that .
0	51	my body felt normal except it was different .
0	52	i opened my eyes and looked over to where i just had been or still was .
0	53	my human body slightly glowed as it sat in meditation under the tree .
0	54	i could see the monster clearly now that i had stepped into its world .
0	55	it was hunched over , occasionally touching the ground with one of its curled front paws .
0	56	dark saliva dripped out of its mouth .
0	57	the beady black eyes made it no less scary .
0	58	i knew exactly what it is .
0	59	a shiver ran down my spine .
0	60	thrashers were probably one of the most feral and nasty monsters ive ever encountered .
0	61	they sometimes travel alone or hunt in packs .
0	62	theyre awful in packs , extremely hard to take out without their knowledge of your presence .
0	63	i placed an arrow in my bow , and aimed for the head .
0	64	i couldnt let the creature reach my house .
0	65	i pulled back the string , and then stopped .
0	66	i couldnt take out one of these like , this unless i set it on fire .
0	67	i could take it out with a few arrows but its best to kill thrashers before they realize youre there .
0	68	instead i lowered my bow and slowly crept up to it , until i could see the hideous claws on its hands and the hatred in its eyes .
0	69	thrashers hate everything and sometimes i wonder how they tolerate each other .
0	70	i aimed for the head again and released the arrow .
0	71	it whizzed silently into the thrasher and it howled in pain as the arrow dug into its fur .
0	72	it whirled around , facing me .
0	73	a low growl sounded in its throat and it charged .
0	74	i ducked as it slashed my side and i winced .
0	75	my shirt slowly turned blue from the trickling wound .
0	76	yes , blue , my blood is blue .
0	77	it turned back around and threw its self at me .
0	78	this time i was prepared .
0	79	i pulled out my hunting knife out just as he landed on me .
0	80	he yelped and went limp .
0	81	i pushed him off and grabbed my fallen bow , the familiar feeling of the wood boosting my confidence .
0	82	i turned back around and started back where i left off , but then a body slammed into me .
0	83	i just had enough time to see the thrasher , before the fur and dirt filled my vision .
0	84	i spit out the clump of wiry hair that had somehow found its way into my gasping mouth .
0	85	i felt a searing pain as the thrasher raked my left leg .
0	86	i closed my eyes against the pain , trying to focus myself against the dull throb in my shin .
0	87	then the thrasher went limp against me .
0	88	it was hauled off of my body and i found myself looking up into the face of a very unhappy dark blue haired boy .
0	89	his sword was gripped tightly in his clenched fingers and a look of amusement , or many it was anger , shone on his face .
0	90	i swallowed and collected my weapons from the ground , not daring to look at my leg .
0	91	hi seth , this is unexpected .
0	92	seth glared at me and pulled me up .
0	93	he was definitely angry at me .
0	94	i tried to place weight on my leg but winced as pain shot up it .
0	95	he kicked the dead thrasher one more time before sheathing his sword .
0	96	you know better jazell then to take on a monster by yourself .
0	97	you never know what kind it is until you have already changed , he scolded me and i could feel a lecture coming .
0	98	i looked down at leg .
0	99	claw marks ran up my shin and my blue blood was soaking my pants .

End of preview. Expand in Data Studio

BookCorpus — Cleaned for Pre-training LLMs

A cleaned, deduplicated, document-segmented version of SamuelYang/bookcorpus

TL;DR

Property	Value
Rows (sentences)	33,649,142
Documents (books)	4,086
Format	CSV, 3 columns: `doc_id`, `sent_id`, `text`
Language	English (lowercased)
Source	`SamuelYang/bookcorpus` (74,004,228 raw rows)

Schema

Column	Type	Description
`doc_id`	int	Inferred document/book identifier. Sentences sharing the same `doc_id` come from the same book.
`sent_id`	int	Sentence position within its document (0-indexed). Preserves original order.
`text`	string	Cleaned sentence text (lowercased, normalized).

How to use it

Quick load

from datasets import load_dataset

ds = load_dataset("kd13/bookcorpus-clean", split="train")
print(ds[0])
# {'doc_id': 0, 'sent_id': 0, 'text': 'i wish i had a better answer ...'}

Cleaning pipeline

Applied in this order to the source dataset:

Unicode + whitespace normalization — NFKC normalization, collapse consecutive whitespace, strip.
Document segmentation — since the source is a flat stream of sentences without book IDs, document boundaries are inferred from telltale markers at the start of books:
- ISBN lines (e.g. isbn : 1492913731)
- Copyright declarations (copyright 2013 ...)
- all rights reserved
- chapter 1
Line-level filters — sentences are dropped if they:
- have fewer than 20 or more than 1000 characters
- match boilerplate patterns (copyright, ISBN, "all rights reserved")
- have an alphabetic-character ratio below 0.6
- have a digit ratio above 0.3
- contain no alphabetic characters
Language filter — cheap English stop-word ratio check (≥ 5% of tokens must be in a small English stop-word set; short lines pass through).
Within-document exact dedup — SHA-1 hashing drops repeated sentences inside the same book (e.g. recurring chapter headers, section dividers). Note: dedup is not applied globally — sentences like "he nodded." occur legitimately across many books.
Document filter — books with fewer than 8 surviving sentences are dropped (not enough context for NSP).
Cross-document near-duplicate removal — a SHA-1 fingerprint of each document's first 5 sentences identifies same-book re-uploads; duplicates are dropped.

Cleaning statistics

Metric	Value
Raw rows (sentences) in source	74,004,228
Documents detected	6,779
Documents kept	4,086
Documents dropped (< 8 sentences)	973
Documents dropped (near-duplicate)	1,720
Sentences kept	33,649,142

Drop rate: ~40% of detected documents removed (mostly same-book re-uploads and too-short documents).

Source & licensing

Source dataset: SamuelYang/bookcorpus
Original corpus: BookCorpus (Zhu et al., 2015), originally scraped from Smashwords. The original BookCorpus has well-documented provenance and consent concerns; downstream users should review them before commercial use.
This cleaned derivative is released under the MIT License for the cleaning code and structuring effort. The underlying text retains whatever rights apply to the upstream source.

Downloads last month: 24