#2541 | AsCollected

#2541 | AsCollected

AsCollected #2541 -

'NANOCHAT-FILIPINO P6-M'

https://ascollected.org/XZ8_TI5

Authors (anonymous for now)

Author 1

Dates
Made Public: 2026-08-25
Latest update: 2026-08-25

✕ Introduction

Every AsCollected includes two tables; Table 1 shows author roles, Table 2 data provenance.
Authors are able to edit those tables at any point
Changes made after a project is public are done with tracked-changes on
This project is public but currently anonymous (enabling blind peer-review).

Table 1: Author roles

	DATA				CODE
	Raw		Final		Data Cleaning			Data Analysis
	obtained	has copy	has copy		wrote	reviewed	has copy	wrote	reviewed	has copy
	Datasources [s_id: #3781] Public data [d_id: #4276]	1	1	1		1	--	1	1	--	1

Table 2: Data collected

	When	Where	Data ID	Non-authors	Additional Info
Datasources [s_id: #3781] Public data [d_id: #4276]	Obtained 2026-08-24 1	Obtained from https://huggingface.co/datasets/Salesforce/wikitext & https://huggingface.co/datasets/linkanjarad/Wikitext-TL39 1	Unique data ID? Yes HF wikitext-103-raw-v1/train-00000-of-00002.parquet (+ sibling shards); frozen study files english_{train,val,test}.jsonl and HF parquet identity + frozen (read more) Unique data ID? Yes HF wikitext-103-raw-v1/train-00000-of-00002.parquet (+ sibling shards); frozen study files english_{train,val,test}.jsonl and HF parquet identity + frozen train.jsonl / val.jsonl / test.jsonl 1	RAs involved No 1	Anything else: https://huggingface.co/datasets/Salesforce/wikitext · config wikitext-103-raw-v1 · revision b08601e04326c79dfdd32d625aee71d232d685c3 2026-08-24 (P6 Gate B ver (read more) Anything else: https://huggingface.co/datasets/Salesforce/wikitext · config wikitext-103-raw-v1 · revision b08601e04326c79dfdd32d625aee71d232d685c3 2026-08-24 (P6 Gate B verify/bind; family originally frozen in P2) Official EN train/val/test JSONL SHA-256: train 09ae691caebb33a4bb81db4e570f630cac9ede11cb4116b2e08a3dbe08ef775a; val 874dec29844b3d46fc39e5479ee2dc4b3ba37309d9baf3bba4b5654697f3ae3b; test 2bccabc020cbb8d09273cccdc42ed926957b83824ca767c96fb588041b8d434e. Test text is not redistributed in ResearchBox. Not WikiText-2; not word-level wikitext-103-v1. Unique File Name? Yes — e.g. HF wikitext-103-raw-v1/train-00000-of-00002.parquet (+ sibling shards); frozen study files english_{train,val,test}.jsonl https://huggingface.co/datasets/linkanjarad/Wikitext-TL39 (Cruz & Cheng WikiText-TL-39 lineage) 2026-08-24 (P6 Gate B; P1.1-eligible split reuse) Frozen TL train/val/test SHA-256: train 2b0474c5700dc1eba14def572aa23cc227e4c59c10c2de3ce6b7bda75d137687; val 4d51644b84d05050bfc8c515079e60f6e437082b6cce2122e9ed00e7b1db2b1c; test 3bd193458f4c494d84dae345548c0c01cb6cd7275e98d6ed39a41d517a093baf. Holdout test JSONL not deposited. Does not amend P1.1–P5 AsCollected records. Unique File Name? Yes — HF parquet identity + frozen train.jsonl / val.jsonl / test.jsonl Data cleaned: With code 1

Notes:
Numbers on top-right of cell represent the author(s) providing the information.