diff --git "a/LICENSE_AUDIT.md" "b/LICENSE_AUDIT.md" deleted file mode 100644--- "a/LICENSE_AUDIT.md" +++ /dev/null @@ -1,1536 +0,0 @@ -# Source license review — 2026-09-30 - -This review covers all 129 previously recorded source URLs and two additional original-source references, across the 182 retained configurations. It checks published license declarations and their scope, not legal clearance. A license on code, a wrapper, metadata, or annotations does not automatically license underlying documents. Acquisition revisions remain unchanged; reviewed live pages are identified by their evidence URLs. - -`verified` means an explicit declaration was found for the stated scope; `qualified` means scope limits, multiple terms, or conflicting declarations require the accompanying notes; `unresolved` means an applicable data grant was not established. These categories are evidence descriptions, not approvals for a use case. - -The six S1MB-authored evaluation subsets have no external upstream and retain their `not separately declared` license status. This review does not assign them a license. - - -## https://coliee.org/archive/coliee-2022 - -Source: [https://coliee.org/archive/coliee-2022](https://coliee.org/archive/coliee-2022) - -Recorded terms: **unresolved: COLIEE access/application terms; no public data license verified**. Status: `unresolved`. - -BGE-M3 cites the COLIEE 2022 overview, but does not identify an exact task/release license. A public data reuse/redistribution grant could not be verified. The license on an overview paper is not a license on its dataset. - -- [Collected-data appendix lists COLIEE and cites the COLIEE 2022 summary.](https://arxiv.org/html/2402.03216v4) -- [Official edition page directs users seeking previous data to contact the organizers.](https://coliee.org/archive/coliee-2022) -- [Current official access page requires application forms for past datasets.](https://coliee.org/data-request) - - -## PolyAI-LDN/task-specific-datasets - -Source: [https://github.com/PolyAI-LDN/task-specific-datasets](https://github.com/PolyAI-LDN/task-specific-datasets) - -Recorded terms: **cc-by-4.0**. Status: `verified`. - -Repository README directs users to LICENSE; pinned LICENSE says Attribution 4.0 International, covering datasets shared in repository. - -- [README states that datasets shared in the repository are licensed under the LICENSE file.](https://github.com/PolyAI-LDN/task-specific-datasets/blob/57ec275d8078af65b7731c2a98be812d844a6d6b/README.md) -- [Fetched pinned LICENSE; its opening line says “Attribution 4.0 International”.](https://github.com/PolyAI-LDN/task-specific-datasets/blob/57ec275d8078af65b7731c2a98be812d844a6d6b/LICENSE) - - -## UKPLab/acl2022-impli - -Source: [https://github.com/UKPLab/acl2022-impli](https://github.com/UKPLab/acl2022-impli) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README has no license declaration; this page alone does not substantiate a license.](https://github.com/UKPLab/acl2022-impli/raw/HEAD/README.md) - - -## causalNLP/cladder - -Source: [https://github.com/causalNLP/cladder](https://github.com/causalNLP/cladder) - -Recorded terms: **mit**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license mit; retained card text is in the evidence snapshot.](https://github.com/causalNLP/cladder/blob/3d2d1169b4b939a09048a6a75956c8972a93cc38/README.md) -- [Repository LICENSE file declares MIT. This is the project license; verify any separately sourced material remains subject to its own terms.](https://github.com/causalNLP/cladder/blob/master/LICENSE) - - -## google-research-datasets/Taskmaster - -Source: [https://github.com/google-research-datasets/Taskmaster](https://github.com/google-research-datasets/Taskmaster) - -Recorded terms: **cc-by-4.0**. Status: `verified`. - -Official TM-1, TM-2, TM-3, and TM-4 subdirectory READMEs each include a COPYRIGHT NOTICE stating the work is made available under Creative Commons Attribution 4.0. The top-level repository lacks a license file, but the source subdirectories cover the collection. Snapshots: taskmaster-TM-*-README.md. - -- [COPYRIGHT NOTICE says the work is made available under Creative Commons Attribution 4.0 License.](https://github.com/google-research-datasets/Taskmaster/blob/master/TM-1-2019/README.md) -- [COPYRIGHT NOTICE says the work is made available under Creative Commons Attribution 4.0 License.](https://github.com/google-research-datasets/Taskmaster/blob/master/TM-2-2020/README.md) -- [COPYRIGHT NOTICE says the work is made available under Creative Commons Attribution 4.0 License.](https://github.com/google-research-datasets/Taskmaster/blob/master/TM-3-2020/README.md) -- [COPYRIGHT NOTICE says the work is made available under Creative Commons Attribution 4.0 License.](https://github.com/google-research-datasets/Taskmaster/blob/master/TM-4-2024/README.md) - - -## google-research-datasets/dstc8-schema-guided-dialogue - -Source: [https://github.com/google-research-datasets/dstc8-schema-guided-dialogue](https://github.com/google-research-datasets/dstc8-schema-guided-dialogue) - -Recorded terms: **cc-by-sa-4.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README has no license declaration; this page alone does not substantiate a license.](https://github.com/google-research-datasets/dstc8-schema-guided-dialogue/raw/HEAD/README.md) - - -## nyu-mll/crows-pairs - -Source: [https://github.com/nyu-mll/crows-pairs](https://github.com/nyu-mll/crows-pairs) - -Recorded terms: **cc-by-sa-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-sa-4.0; retained card text is in the evidence snapshot.](https://github.com/nyu-mll/crows-pairs/blob/8aaac11c485473159ec9328a65253a5be9a479dc/README.md) - - -## princeton-nlp/SimCSE - -Source: [https://github.com/princeton-nlp/SimCSE](https://github.com/princeton-nlp/SimCSE) - -Recorded terms: **SNLI: CC-BY-SA-4.0; MultiNLI: OANC terms, CC-BY-3.0, CC-BY-SA-3.0 and public-domain portions**. Status: `qualified`. - -BGE-M3 identifies its NLI source as SimCSE. SimCSE uses SNLI and MultiNLI for supervised training. These are corpus terms, not the MIT license on SimCSE code. MultiNLI fiction terms vary by work and jurisdiction; the original paper describes the Share-Alike 3.0 grant. - -- [The English fine-tuning source table identifies NLI from SimCSE.](https://huggingface.co/datasets/Shitao/bge-m3-data/blob/main/README.md) -- [The supervised training data are SNLI and MNLI.](https://github.com/princeton-nlp/SimCSE#training) -- [The corpus release is identified as CC-BY-SA-4.0.](https://nlp.stanford.edu/projects/snli/) -- [Distribution section describes OANC terms, Share-Alike 3.0 for Seven Swords, CC-BY-3.0 for two named works, and US public-domain fiction.](https://cims.nyu.edu/~sbowman/multinli/paper.pdf) - - -## sonos/nlu-benchmark - -Source: [https://github.com/sonos/nlu-benchmark](https://github.com/sonos/nlu-benchmark) - -Recorded terms: **cc0-1.0**. Status: `verified`. - -Pinned repository LICENSE explicitly says CC0 1.0 Universal. - -- [Fetched pinned LICENSE; its opening line says “CC0 1.0 Universal”.](https://github.com/sonos/nlu-benchmark/blob/b86ac7f1577868c42158d0dec77db50956046696/LICENSE) - - -## stanfordnlp/contract-nli - -Source: [https://github.com/stanfordnlp/contract-nli](https://github.com/stanfordnlp/contract-nli) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README has no license declaration; this page alone does not substantiate a license.](https://github.com/stanfordnlp/contract-nli/raw/HEAD/README.md) - - -## uclanlp/corefBias - -Source: [https://github.com/uclanlp/corefBias](https://github.com/uclanlp/corefBias) - -Recorded terms: **mit**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license mit; retained card text is in the evidence snapshot.](https://github.com/uclanlp/corefBias/blob/0bce984dd081bbc10b0622f326727a024c607895/README.md) -- [Repository LICENSE file declares MIT. This is the project license; verify any separately sourced material remains subject to its own terms.](https://github.com/uclanlp/corefBias/blob/master/LICENSE) - - -## AmazonScience/massive - -Source: [https://huggingface.co/datasets/AmazonScience/massive](https://huggingface.co/datasets/AmazonScience/massive) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['cc-by-4.0']; the card body/context is preserved in 04-README.md.](https://huggingface.co/datasets/AmazonScience/massive/blob/ff6bd8e4b27c3543e4f8fe2108f32bb95a6f8740/README.md) - - -## Anthropic/hh-rlhf - -Source: [https://huggingface.co/datasets/Anthropic/hh-rlhf](https://huggingface.co/datasets/Anthropic/hh-rlhf) - -Recorded terms: **mit**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: mit.](https://huggingface.co/datasets/Anthropic/hh-rlhf/raw/main/README.md) - - -## DFKI-SLT/few-nerd - -Source: [https://huggingface.co/datasets/DFKI-SLT/few-nerd](https://huggingface.co/datasets/DFKI-SLT/few-nerd) - -Recorded terms: **cc-by-sa-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-sa-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/DFKI-SLT/few-nerd/blob/205f3e9c9f3577ea2561d43f2f62dc249ab92d5b/README.md) - - -## FastFit/hwu_64 - -Source: [https://huggingface.co/datasets/FastFit/hwu_64](https://huggingface.co/datasets/FastFit/hwu_64) - -Recorded terms: **cc-by-4.0**. Status: `verified`. - -Original data-author repository identifies the collected home-domain NLU dataset (25,716 annotated utterances, 64 intents) and explicitly states all data are released under CC BY 4.0. Some benchmark repackagings report CC BY-SA 3.0, so cite the direct source-author declaration and preserve this discrepancy for derivative distributions. Snapshot: evidence/batch-1/hwu-original-readme.txt. - -- [Original author README describes collected home-domain human-robot interaction data and says all data are released under Creative Commons Attribution 4.0 International (CC BY 4.0); it reports 25,716 annotated lines.](https://github.com/xliuhw/NLU-Evaluation-Data/blob/master/README.md) -- [Repository LICENSE is the Creative Commons Attribution 4.0 legal code.](https://github.com/xliuhw/NLU-Evaluation-Data/blob/master/LICENSE) - - -## HuggingFaceFW/fineweb-edu - -Source: [https://huggingface.co/datasets/HuggingFaceFW/fineweb-edu](https://huggingface.co/datasets/HuggingFaceFW/fineweb-edu) - -Recorded terms: **odc-by**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: odc-by.](https://huggingface.co/datasets/HuggingFaceFW/fineweb-edu/raw/main/README.md) - - -## LocalLLaMA/typed-decisions - -Source: [https://huggingface.co/datasets/LocalLLaMA/typed-decisions](https://huggingface.co/datasets/LocalLLaMA/typed-decisions) - -Recorded terms: **apache-2.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license apache-2.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/LocalLLaMA/typed-decisions/blob/main/README.md) - - -## OpenAssistant/oasst1 - -Source: [https://huggingface.co/datasets/OpenAssistant/oasst1](https://huggingface.co/datasets/OpenAssistant/oasst1) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['apache-2.0']; the card body/context is preserved in 06-README.md.](https://huggingface.co/datasets/OpenAssistant/oasst1/blob/fdf72ae0827c1cda404aff25b6603abec9e3399b/README.md) - - -## Paul/hatecheck - -Source: [https://huggingface.co/datasets/Paul/hatecheck](https://huggingface.co/datasets/Paul/hatecheck) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: - cc-by-4.0.](https://huggingface.co/datasets/Paul/hatecheck/raw/main/README.md) - - -## PolyAI/banking77 - -Source: [https://huggingface.co/datasets/PolyAI/banking77](https://huggingface.co/datasets/PolyAI/banking77) - -Recorded terms: **cc-by-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/PolyAI/banking77/blob/90d4e2ee5521c04fc1488f065b8b083658768c57/README.md) - - -## PolyAI/minds14 - -Source: [https://huggingface.co/datasets/PolyAI/minds14](https://huggingface.co/datasets/PolyAI/minds14) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['cc-by-4.0']; the card body/context is preserved in 07-README.md.](https://huggingface.co/datasets/PolyAI/minds14/blob/40ce77cb32a384e4d50a568e1ec39ac804019d33/README.md) - - -## Samsoup/cosmos_qa - -Source: [https://huggingface.co/datasets/Samsoup/cosmos_qa](https://huggingface.co/datasets/Samsoup/cosmos_qa) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README has no license declaration; this page alone does not substantiate a license.](https://huggingface.co/datasets/Samsoup/cosmos_qa/raw/main/README.md) - - -## SetFit/enron_spam - -Source: [https://huggingface.co/datasets/SetFit/enron_spam](https://huggingface.co/datasets/SetFit/enron_spam) - -Recorded terms: **unknown (no dataset license located)**. Status: `unresolved`. - -The original AUEB dataset page identifies Enron-Spam as ham messages from Enron plus added spam and offers raw/preprocessed files, but no dataset license was found. A third-party cleaned CSV repository’s GPL-3.0 is a separate repackaging/code claim and is not adopted as the upstream data license. - -- [AUEB official data catalog describes Enron-Spam as Enron ham mail plus spam mail and links the original distribution; no license statement is given.](https://nlp.cs.aueb.gr/en/software_data.html) -- [A third-party repackaging says it processes/downloads original AUEB data; its GPL-3.0 repository license does not establish AUEB upstream dataset licensing.](https://github.com/MWiechmann/enron_spam_data) - - -## SetFit/ethos_binary - -Source: [https://huggingface.co/datasets/SetFit/ethos_binary](https://huggingface.co/datasets/SetFit/ethos_binary) - -Recorded terms: **gpl-3.0**. Status: `qualified`. - -The upstream ETHOS repository contains the binary dataset and its LICENSE says GNU GPL version 3. The upstream license file applies to a repository containing software and dataset; the HF conversion itself has no license statement. Review scope before relying on GPL as a data license. - -- [Repository README describes Ethos_Dataset_Binary.csv as part of the repository dataset.](https://github.com/intelligence-csd-auth-gr/Ethos-Hate-Speech-Dataset) -- [Repository LICENSE is GNU General Public License version 3.](https://github.com/intelligence-csd-auth-gr/Ethos-Hate-Speech-Dataset/blob/master/LICENSE) - - -## Shitao/bge-m3-data - -Source: [https://huggingface.co/datasets/Shitao/bge-m3-data](https://huggingface.co/datasets/Shitao/bge-m3-data) - -Recorded terms: **No aggregate license declared; constituent terms apply**. Status: `qualified`. - -The BGE-M3 card enumerates its constituent datasets but declares no aggregate data license. Original-source terms are listed separately for the retained configurations in README. This review adds the previously missing COLIEE and SimCSE/SNLI/MultiNLI references. COLIEE reuse/redistribution terms remain unresolved. A model or training-code MIT license does not replace these corpus terms. - -- [Card lists constituent sources including MS MARCO, NQ, HotpotQA, TriviaQA, SQuAD, COLIEE, PubMedQA, NLI from SimCSE, Mr.TyDi, and MIRACL; no license is declared.](https://huggingface.co/datasets/Shitao/bge-m3-data/raw/main/README.md) - - -## ZefanCai/Open-Jev - -Source: [https://huggingface.co/datasets/ZefanCai/Open-Jev](https://huggingface.co/datasets/ZefanCai/Open-Jev) - -Recorded terms: **CC0-1.0; Generated conversations: CC0-1.0; upstream question descriptions: TypeSafe documentation, license not verified; CC0-1.0 (generated numeric trajectories; no game assets)**. Status: `qualified`. - -The card explicitly limits CC0 to generated records; some included content is derived from TypeSafe documentation or external source material with separate or unverified terms. Treat the aggregate dataset as mixed; CC0 applies only to identified generated subsets. - -- [Dataset card lists record/config-specific provenance and CC0-1.0 for generated material, while expressly excluding upstream wording and external assets from that dedication.](https://huggingface.co/datasets/ZefanCai/Open-Jev/blob/c67699e13d0ae25e35b77165a4b6b079bedc8aba/README.md) - - -## allenai/ai2_arc - -Source: [https://huggingface.co/datasets/allenai/ai2_arc](https://huggingface.co/datasets/allenai/ai2_arc) - -Recorded terms: **cc-by-sa-4.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['cc-by-sa-4.0']; the card body/context is preserved in 09-README.md.](https://huggingface.co/datasets/allenai/ai2_arc/blob/210d026faf9955653af8916fad021475a3f00453/README.md) - - -## allenai/art - -Source: [https://huggingface.co/datasets/allenai/art](https://huggingface.co/datasets/allenai/art) - -Recorded terms: **unknown**. Status: `unresolved`. - -The original repository offers an Apache-2.0 LICENSE, but the separate official data archives anli.zip and anlg.zip contain only JSONL data/labels and no README, LICENSE, NOTICE, or COPYING file. Thus the code license does not establish dataset terms; no data-specific grant was found in the packaged downloads. - -- [Live HF dataset card frontmatter explicitly sets license: unknown and points to the original repository.](https://huggingface.co/datasets/allenai/art/raw/main/README.md) -- [Original repository has an Apache-2.0 license file, but it does not explicitly describe the separately distributed dataset archives.](https://github.com/allenai/abductive-commonsense-reasoning/blob/master/LICENSE) -- [Inspected official 5.4 MB dataset archive: contains train/dev/test JSONL and label files only; no README/license/notice file.](https://storage.googleapis.com/ai2-mosaic/public/abductive-commonsense-reasoning-iclr2020/anli.zip) -- [Inspected official 31.5 MB dataset archive: contains train/dev/test JSONL files only; no README/license/notice file.](https://storage.googleapis.com/ai2-mosaic/public/abductive-commonsense-reasoning-iclr2020/anlg.zip) - - -## allenai/gooaq - -Source: [https://huggingface.co/datasets/allenai/gooaq](https://huggingface.co/datasets/allenai/gooaq) - -Recorded terms: **apache-2.0 (repository LICENSE); non-commercial research restriction (dataset README)**. Status: `qualified`. - -GooAQ README explicitly says “This dataset should not be used for any commercial purposes” and points to the LICENSE. The repository LICENSE is Apache-2.0, creating an apparent mismatch; record the dataset restriction and do not present the repository license as an unambiguous commercial data grant. - -- [README under the project overview says the dataset should not be used for commercial purposes and refers to the license.](https://github.com/allenai/gooaq) -- [Repository LICENSE is Apache-2.0; it conflicts with the README dataset-specific non-commercial warning.](https://github.com/allenai/gooaq/blob/main/LICENSE) -- [Open issue records the discrepancy between Apache-2.0 LICENSE and README non-commercial restriction.](https://github.com/allenai/gooaq/issues/4) - - -## allenai/openbookqa - -Source: [https://huggingface.co/datasets/allenai/openbookqa](https://huggingface.co/datasets/allenai/openbookqa) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -HF card frontmatter explicitly says unknown. The original AllenAI repository has an Apache-2.0 LICENSE, but it is a repository-level grant; the dataset page’s card does not itself state a data license. Cite as upstream repository license, qualified for dataset scope. - -- [HF card frontmatter says license: unknown.](https://huggingface.co/datasets/allenai/openbookqa) -- [Original repository LICENSE is Apache License 2.0.](https://github.com/allenai/OpenBookQA/blob/master/LICENSE) - - -## allenai/qasc - -Source: [https://huggingface.co/datasets/allenai/qasc](https://huggingface.co/datasets/allenai/qasc) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: - cc-by-4.0.](https://huggingface.co/datasets/allenai/qasc/raw/main/README.md) - - -## allenai/qasper - -Source: [https://huggingface.co/datasets/allenai/qasper](https://huggingface.co/datasets/allenai/qasper) - -Recorded terms: **cc-by-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/allenai/qasper/blob/fdc9d8214fbab5dd782958601db4d678e6934a54/README.md) - - -## allenai/quartz - -Source: [https://huggingface.co/datasets/allenai/quartz](https://huggingface.co/datasets/allenai/quartz) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['cc-by-4.0']; the card body/context is preserved in 11-README.md.](https://huggingface.co/datasets/allenai/quartz/blob/28c1dbb56caf81799296cb17892fa73402e23464/README.md) - - -## allenai/ropes - -Source: [https://huggingface.co/datasets/allenai/ropes](https://huggingface.co/datasets/allenai/ropes) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: - cc-by-4.0.](https://huggingface.co/datasets/allenai/ropes/raw/main/README.md) - - -## allenai/scitail - -Source: [https://huggingface.co/datasets/allenai/scitail](https://huggingface.co/datasets/allenai/scitail) - -Recorded terms: **apache-2.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license apache-2.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/allenai/scitail/blob/0cc4353235b289165dfde1c7c5d1be983f99ce44/README.md) - - -## allenai/winogrande - -Source: [https://huggingface.co/datasets/allenai/winogrande](https://huggingface.co/datasets/allenai/winogrande) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -The HF card has no license field. The original AllenAI repository includes an Apache-2.0 LICENSE. This is repository-level evidence; retain scope caveat for dataset content. - -- [Original WinoGrande repository LICENSE is Apache License 2.0.](https://github.com/allenai/winogrande/blob/master/LICENSE) - - -## allenai/wiqa - -Source: [https://huggingface.co/datasets/allenai/wiqa](https://huggingface.co/datasets/allenai/wiqa) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README has no license declaration; this page alone does not substantiate a license.](https://huggingface.co/datasets/allenai/wiqa/raw/main/README.md) - - -## amydeng2000/CREAK - -Source: [https://huggingface.co/datasets/amydeng2000/CREAK](https://huggingface.co/datasets/amydeng2000/CREAK) - -Recorded terms: **MIT (repository license; data grant not explicit)**. Status: `qualified`. - -The HF card has no license statement. Original CREAK repository LICENSE.txt is the MIT license but defines the licensed work as “Software” and associated documentation; it does not expressly identify the dataset files as licensed. Keep the data grant qualified. - -- [Original repository README identifies the CREAK data files under data/creak.](https://github.com/yasumasaonoe/creak) -- [LICENSE.txt is MIT, but its grant is framed as rights to the “Software” and associated documentation; it does not expressly name dataset data files.](https://github.com/yasumasaonoe/creak/blob/master/LICENSE.txt) -- [The source card links to the original CREAK repository and contains no dataset license declaration.](https://huggingface.co/datasets/amydeng2000/CREAK) - - -## aps/super_glue - -Source: [https://huggingface.co/datasets/aps/super_glue](https://huggingface.co/datasets/aps/super_glue) - -Recorded terms: **bsd-2-clause (COPA config)**. Status: `verified`. - -This source row uses only SuperGLUE’s COPA config. SuperGLUE says tasks derive from individual source datasets; the official COPA project page has a dedicated License section reproducing the BSD 2-Clause license and USC copyright notice. - -- [Official COPA page has section “9. License (BSD 2-Clause License)” and reproduces BSD-2-Clause redistribution conditions and the USC copyright notice.](https://asgordon.github.io/copa.html#license-bsd-2-clause-license) -- [SuperGLUE says tasks derive from existing datasets and asks users to consult each original license; COPA’s dedicated source page supplies BSD-2-Clause terms.](https://huggingface.co/datasets/aps/super_glue/blob/3de24cf8022e94f4ee4b9d55a6f539891524d646/README.md) - - -## baber/social_i_qa - -Source: [https://huggingface.co/datasets/baber/social_i_qa](https://huggingface.co/datasets/baber/social_i_qa) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README has no license declaration; this page alone does not substantiate a license.](https://huggingface.co/datasets/baber/social_i_qa/raw/main/README.md) - - -## bitext/Bitext-customer-support-llm-chatbot-training-dataset - -Source: [https://huggingface.co/datasets/bitext/Bitext-customer-support-llm-chatbot-training-dataset](https://huggingface.co/datasets/bitext/Bitext-customer-support-llm-chatbot-training-dataset) - -Recorded terms: **cdla-sharing-1.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cdla-sharing-1.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/bitext/Bitext-customer-support-llm-chatbot-training-dataset/blob/430d1a89bd93bd1fa23c16f29dd53e73f0087443/README.md) - - -## castorini/mr-tydi - -Source: [https://huggingface.co/datasets/castorini/mr-tydi](https://huggingface.co/datasets/castorini/mr-tydi) - -Recorded terms: **apache-2.0 (Mr. TyDi release); underlying Wikipedia text terms remain applicable**. Status: `qualified`. - -Mr. TyDi official repository explicitly says dataset is Apache-2.0. The retained English collection is built on TyDi; official TyDi QA documentation says source articles are Wikipedia snapshots. Apache release metadata should not be interpreted as overriding applicable Wikipedia text licenses; preserve attribution/share-alike obligations for source text. - -- [Official Mr. TyDi README says “Mr. TyDi is licensed under the Apache License 2.0.”](https://github.com/castorini/mr.tydi) -- [Official TyDi QA source-data section says articles are drawn from coherent Wikipedia snapshots and links archived dumps.](https://github.com/google-research-datasets/tydiqa#source-data) -- [Wikimedia terms establish CC BY-SA and GFDL terms for text; the Wikipedia source text is not automatically relicensed by the derived package.](https://foundation.wikimedia.org/wiki/Terms_of_Use?useformat=mobile) - - -## clinc/clinc_oos - -Source: [https://huggingface.co/datasets/clinc/clinc_oos](https://huggingface.co/datasets/clinc/clinc_oos) - -Recorded terms: **cc-by-3.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: - cc-by-3.0.](https://huggingface.co/datasets/clinc/clinc_oos/raw/main/README.md) - - -## deepmind/aqua_rat - -Source: [https://huggingface.co/datasets/deepmind/aqua_rat](https://huggingface.co/datasets/deepmind/aqua_rat) - -Recorded terms: **apache-2.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license apache-2.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/deepmind/aqua_rat/blob/33301c6a050c96af81f63cad5562cb5363e88971/README.md) - - -## fancyzhx/ag_news - -Source: [https://huggingface.co/datasets/fancyzhx/ag_news](https://huggingface.co/datasets/fancyzhx/ag_news) - -Recorded terms: **unknown**. Status: `unresolved`. - -Reviewed original AG News corpus homepage and HF card; no license grant located. Homepage describes corpus and source news material; no declaration is not permission to reuse. - -- [Original AG News corpus page provides corpus description/download context but no license grant was located.](http://groups.di.unipi.it/~gulli/AG_corpus_of_news_articles.html) - - -## fancyzhx/dbpedia_14 - -Source: [https://huggingface.co/datasets/fancyzhx/dbpedia_14](https://huggingface.co/datasets/fancyzhx/dbpedia_14) - -Recorded terms: **cc-by-sa-3.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: - cc-by-sa-3.0.](https://huggingface.co/datasets/fancyzhx/dbpedia_14/raw/main/README.md) - - -## google-research-datasets/circa - -Source: [https://huggingface.co/datasets/google-research-datasets/circa](https://huggingface.co/datasets/google-research-datasets/circa) - -Recorded terms: **cc-by-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/google-research-datasets/circa/blob/faa1b5a78dd926a899bcd4da289c2e3abe8061a9/README.md) - - -## google-research-datasets/go_emotions - -Source: [https://huggingface.co/datasets/google-research-datasets/go_emotions](https://huggingface.co/datasets/google-research-datasets/go_emotions) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['apache-2.0']; the card body/context is preserved in 16-README.md.](https://huggingface.co/datasets/google-research-datasets/go_emotions/blob/add492243ff905527e67aeb8b80c082af02207c3/README.md) - - -## google-research-datasets/natural_questions - -Source: [https://huggingface.co/datasets/google-research-datasets/natural_questions](https://huggingface.co/datasets/google-research-datasets/natural_questions) - -Recorded terms: **cc-by-sa-3.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: - cc-by-sa-3.0.](https://huggingface.co/datasets/google-research-datasets/natural_questions/raw/main/README.md) - - -## google-research-datasets/paws - -Source: [https://huggingface.co/datasets/google-research-datasets/paws](https://huggingface.co/datasets/google-research-datasets/paws) - -Recorded terms: **PAWS dataset: Google custom grant (“may be freely used for any purpose”); PAWS-Wiki source text: CC BY-SA 3.0 / GFDL source terms; PAWS-QQP excluded due to QQP terms**. Status: `qualified`. - -Local release adapter selects PAWS-Wiki labeled_final only. The original PAWS LICENSE grants free use of the dataset, while its Wiki sentences originate from Wikipedia under the 2019-era CC BY-SA 3.0/GFDL terms. These permissions should be represented separately. PAWS-QQP is not part of this local use; its README says it cannot be directly distributed due to QQP licensing. - -- [Original PAWS LICENSE says the dataset may be freely used for any purpose and requests acknowledgment of Google; it is a custom grant, not CC BY-SA 4.0.](https://github.com/google-research-datasets/paws/blob/master/LICENSE) -- [README describes PAWS-Wiki as generated from Wikipedia pages and states PAWS-QQP cannot be directly distributed due to QQP licensing.](https://github.com/google-research-datasets/paws/blob/master/README.md) -- [Wikimedia licensing update documents CC BY-SA 3.0 Unported as the license added to Wikimedia text in the relevant era; PAWS-Wiki sourced Wikipedia text must retain applicable original source licensing.](https://meta.wikimedia.org/wiki/Licensing_update/Implementation) -- [Original mapping documentation identifies the Wikipedia source pages/mapping used to construct PAWS-Wiki.](https://github.com/google-research-datasets/paws/blob/master/wiki_raw_and_mapping/README.md) -- [Card configs include labeled_final/labeled_swap/unlabeled_final; repository adapter limits local release to PAWS-Wiki labeled_final.](https://huggingface.co/datasets/google-research-datasets/paws) - - -## google-research-datasets/poem_sentiment - -Source: [https://huggingface.co/datasets/google-research-datasets/poem_sentiment](https://huggingface.co/datasets/google-research-datasets/poem_sentiment) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['cc-by-4.0']; the card body/context is preserved in 17-README.md.](https://huggingface.co/datasets/google-research-datasets/poem_sentiment/blob/685b95a2787a869b7bae6c4480810f57fe23b48e/README.md) - - -## google/boolq - -Source: [https://huggingface.co/datasets/google/boolq](https://huggingface.co/datasets/google/boolq) - -Recorded terms: **cc-by-sa-3.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: - cc-by-sa-3.0.](https://huggingface.co/datasets/google/boolq/raw/main/README.md) - - -## google/civil_comments - -Source: [https://huggingface.co/datasets/google/civil_comments](https://huggingface.co/datasets/google/civil_comments) - -Recorded terms: **cc0-1.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc0-1.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/google/civil_comments/blob/f2970eb3a55777454c94069077cc8d9b5866312d/README.md) - - -## gretelai/gretel-pii-masking-en-v1 - -Source: [https://huggingface.co/datasets/gretelai/gretel-pii-masking-en-v1](https://huggingface.co/datasets/gretelai/gretel-pii-masking-en-v1) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['apache-2.0']; the card body/context is preserved in 18-README.md.](https://huggingface.co/datasets/gretelai/gretel-pii-masking-en-v1/blob/e06eb1499ca8d54470f085021cd8e54f9efac7fd/README.md) - - -## hakari-bench/NanoBEIR-en - -Source: [https://huggingface.co/datasets/hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) - -Recorded terms: **mixed by component (see components)**. Status: `qualified`. - -Active subset mapping now includes primary-source checks for ArguAna, SciDocs, FiQA, and NFCorpus. FiQA has an explicit non-commercial-use restriction and third-party Stack Exchange content licensed per post date. NFCorpus original project/archive provides no explicit license; the later BEIR card claims CC BY-SA 4.0 but does not resolve original-rights provenance. - -- [Card describes NanoBEIR as derived and requires compliance with each upstream dataset license/terms.](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en/raw/main/README.md) - -| Component | Recorded terms | Evidence | -|---|---|---| -| NanoArguAna | cc-by-4.0 | [Original Webis dataset page publishes its license field as CC BY 4.0; Zenodo record 3973258 is its linked dataset distribution.](https://webis.de/data/arguana-counterargs.html)
[Original Webis ArguAna Counterargs dataset record and download.](https://zenodo.org/records/3973258) | -| NanoDBPedia | cc-by-sa-3.0; gfdl | [DBpedia Association states DBpedia data is CC BY-SA 3.0 and GFDL.](https://www.dbpedia.org/about/) | -| NanoFiQA2018 | non-commercial-use-only; cc-by-sa-2.5; cc-by-sa-3.0 | [Original FiQA challenge page states Opinion QA training data is available only for non-commercial use and describes it as Investment Stack Exchange posts.](https://sites.google.com/view/fiqa/home)
[Stack Exchange content terms distinguish CC BY-SA 2.5 for contributions before 2011-04-08 and CC BY-SA 3.0 through 2018-05-01; individual post revision dates govern.](https://money.stackexchange.com/help/licensing)
[FiQA paper describes Opinion QA data as Investment Stack Exchange posts collected 2009–2017.](https://aclanthology.org/2018.finnlp-1.4/) | -| NanoHotpotQA | cc-by-sa-4.0 | [Original HotpotQA project states dataset is released under CC BY-SA 4.0.](https://hotpotqa.github.io/) | -| NanoMSMARCO | mit-noncommercial-research | [Official MS MARCO terms restrict the dataset to non-commercial research use.](https://microsoft.github.io/msmarco/)
[Dataset artifact survey lists MIT license for non-commercial research purposes.](https://aclanthology.org/2025.repl4nlp-1.12.pdf) | -| NanoNFCorpus | unknown | [Original NFCorpus project page identifies the resource and source material but gives no dataset license.](https://www.cl.uni-heidelberg.de/statnlpgroup/nfcorpus/)
[Downloaded original BEIR archive (2.45 MB); archive contains corpus, queries, and qrels only, with no README or LICENSE file.](https://public.ukp.informatik.tu-darmstadt.de/thakur/BEIR/datasets/nfcorpus.zip)
[BEIR mirror declares CC BY-SA 4.0, but this is a later redistribution card and not an original author license statement.](https://huggingface.co/datasets/BeIR/nfcorpus) | -| NanoNQ | cc-by-sa-3.0 | [Google states Natural Questions is CC BY-SA 3.0.](https://ai.google.com/research/NaturalQuestions/download) | -| NanoSCIDOCS | cc-by-4.0 | [Original SciDocs repository LICENSE file is Creative Commons Attribution 4.0 International.](https://github.com/allenai/scidocs/blob/master/LICENSE)
[Repository README says repository includes code, link to data, and instructions; raw corpus is separately downloaded from AI2 S3.](https://github.com/allenai/scidocs) | -| NanoTouche2020 | cc-by-4.0 | [Original Webis dataset page identifies Touché 2020 data under CC BY 4.0.](https://webis.de/data/webis-touche-20.html) | - - -## hakari-bench/NanoCoIR - -Source: [https://huggingface.co/datasets/hakari-bench/NanoCoIR](https://huggingface.co/datasets/hakari-bench/NanoCoIR) - -Recorded terms: **component-specific; no blanket license**. Status: `qualified`. - -NanoCoIR is an aggregate. This component map follows the actual CoIR task names in its card. CoIR paper gives task-specific terms; marked non-commercial/source-derived components retain those restrictions. - -- [License section says this is derived and users must follow upstream licenses, terms, and attribution.](https://huggingface.co/datasets/hakari-bench/NanoCoIR) -- [Ethical Considerations section enumerates license terms for APPS, CodeTransOcean, CodeSearchNet, CoSQA, Synthetic Text2SQL, Code-Feedback, and StackOverflow QA.](https://aclanthology.org/2025.acl-long.1072.pdf) - -| Component | Recorded terms | Evidence | -|---|---|---| -| NanoApps | MIT; non-commercial research only | [CoIR paper says APPS is provided under MIT for non-commercial research purposes.](https://aclanthology.org/2025.acl-long.1072.pdf) | -| NanoCodeFeedbackMT | Apache-2.0 | [CoIR paper says Code-Feedback is provided under Apache License 2.0.](https://aclanthology.org/2025.acl-long.1072.pdf) | -| NanoCodeFeedbackST | Apache-2.0 | [CoIR paper says Code-Feedback is provided under Apache License 2.0.](https://aclanthology.org/2025.acl-long.1072.pdf) | -| NanoCodeSearchNet | MIT; non-commercial research only, plus individual source repository licenses | [CoIR paper says CodeSearchNet is MIT for non-commercial research purposes.](https://aclanthology.org/2025.acl-long.1072.pdf)
[CodeSearchNet row schema records original repository and path; source code files originate in GitHub projects and have individual licenses.](https://github.com/github/CodeSearchNet) | -| NanoCodeSearchNetCCR | MIT; non-commercial research only, plus individual source repository licenses | [CodeSearchNet family is reported under MIT for non-commercial research purposes.](https://aclanthology.org/2025.acl-long.1072.pdf)
[The original code source is repo-attributed; individual snippet licenses remain relevant.](https://github.com/github/CodeSearchNet) | -| NanoCodeTransOceanContest | Apache-2.0 | [CoIR paper says CodeTransOcean is provided under Apache License 2.0.](https://aclanthology.org/2025.acl-long.1072.pdf) | -| NanoCodeTransOceanDL | Apache-2.0 | [CoIR paper says CodeTransOcean is provided under Apache License 2.0.](https://aclanthology.org/2025.acl-long.1072.pdf) | -| NanoCosQA | MIT; non-commercial research only | [CoIR paper says CoSQA is provided under MIT for non-commercial research purposes.](https://aclanthology.org/2025.acl-long.1072.pdf) | -| NanoStackOverflowQA | CC BY-SA 3.0 | [CoIR paper specifies StackOverflow QA is provided under CC-BY-SA 3.0.](https://aclanthology.org/2025.acl-long.1072.pdf)
[Stack Overflow user content is licensed under CC BY-SA; specific applicable version depends on post date/source terms.](https://stackoverflow.com/help/licensing) | -| NanoSyntheticText2SQL | Apache-2.0 | [CoIR paper says Synthetic Text2SQL is provided under Apache License 2.0.](https://aclanthology.org/2025.acl-long.1072.pdf) | - - -## hendrycks/ethics - -Source: [https://huggingface.co/datasets/hendrycks/ethics](https://huggingface.co/datasets/hendrycks/ethics) - -Recorded terms: **mit**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['mit']; the card body/context is preserved in 19-README.md.](https://huggingface.co/datasets/hendrycks/ethics/blob/b8b47c589f8bee77175b8648e5497278b68da48a/README.md) - - -## hotchpotch/arxiv-ir-simulated-search-queries - -Source: [https://huggingface.co/datasets/hotchpotch/arxiv-ir-simulated-search-queries](https://huggingface.co/datasets/hotchpotch/arxiv-ir-simulated-search-queries) - -Recorded terms: **cc0-1.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: cc0-1.0.](https://huggingface.co/datasets/hotchpotch/arxiv-ir-simulated-search-queries/raw/main/README.md) - - -## hotchpotch/fineweb-ir-simulated-search-queries - -Source: [https://huggingface.co/datasets/hotchpotch/fineweb-ir-simulated-search-queries](https://huggingface.co/datasets/hotchpotch/fineweb-ir-simulated-search-queries) - -Recorded terms: **ODC-By 1.0 (FineWeb-Edu database); source-page text retains applicable third-party rights/terms**. Status: `qualified`. - -The dataset card identifies a derived dataset from FineWeb-Edu and records upstream ODC-By. FineWeb’s database license does not relicense underlying web pages; Common Crawl terms also disclaim rights in third-party crawled content. No blanket copyright license for all underlying page text is established. - -- [License section says derived from FineWeb-Edu and states upstream card declares ODC-By; directs users to review source terms.](https://huggingface.co/datasets/hotchpotch/fineweb-ir-simulated-search-queries) -- [FineWeb-Edu card declares ODC-By 1.0 for the database.](https://huggingface.co/datasets/HuggingFaceFW/fineweb-edu) -- [Common Crawl Terms of Use describes limited access/use of its service and warns third-party websites/content remain subject to third-party rights; no grant of their copyrights.](https://www.commoncrawl.org/terms-of-use) - - -## hotchpotch/pubmed-abstract-ir-simulated-search-queries - -Source: [https://huggingface.co/datasets/hotchpotch/pubmed-abstract-ir-simulated-search-queries](https://huggingface.co/datasets/hotchpotch/pubmed-abstract-ir-simulated-search-queries) - -Recorded terms: **CC0-1.0 (aggregate declaration); abstract text rights remain source-specific**. Status: `qualified`. - -The dataset card derives from PubMedAbstractsSubset and repeats its declared CC0-1.0, but abstracts may remain copyrighted by publishers/authors. CC0 should not be presented as blanket rights for the source text. - -- [Card says it is derived from slinusc/PubMedAbstractsSubset and repeats the source CC0-1.0 declaration.](https://huggingface.co/datasets/hotchpotch/pubmed-abstract-ir-simulated-search-queries/blob/main/README.md) -- [NLM says it does not claim copyright on PubMed abstracts; publishers or authors may.](https://www.ncbi.nlm.nih.gov/home/about/policies/) - - -## hotchpotch/wikipedia-english-ir-simulated-search-queries - -Source: [https://huggingface.co/datasets/hotchpotch/wikipedia-english-ir-simulated-search-queries](https://huggingface.co/datasets/hotchpotch/wikipedia-english-ir-simulated-search-queries) - -Recorded terms: **cc-by-sa-4.0; gfdl-where-applicable**. Status: `qualified`. - -Wikipedia-derived text carries source-content terms, while generated query strings have separate authorship/provenance. Dataset card links singletongue/wikipedia-paragraphs as source family; content-level exceptions can vary. - -- [Card says corpus documents derive from Wikipedia and states CC BY-SA 4.0 and GFDL where applicable.](https://huggingface.co/datasets/hotchpotch/wikipedia-english-ir-simulated-search-queries/raw/main/README.md) -- [Wikimedia copyright page states most textual content is CC BY-SA 4.0 and GFDL; exceptions may have different terms.](https://www.mediawiki.org/wiki/Project:Copyrights/en) - - -## hotpotqa/hotpot_qa - -Source: [https://huggingface.co/datasets/hotpotqa/hotpot_qa](https://huggingface.co/datasets/hotpotqa/hotpot_qa) - -Recorded terms: **cc-by-sa-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-sa-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/hotpotqa/hotpot_qa/blob/main/README.md) - - -## ibm-research/argument_quality_ranking_30k - -Source: [https://huggingface.co/datasets/ibm-research/argument_quality_ranking_30k](https://huggingface.co/datasets/ibm-research/argument_quality_ranking_30k) - -Recorded terms: **cc-by-3.0 (official IBM HF dataset metadata; card body conflicts)**. Status: `qualified`. - -The dataset is published under IBM Research’s official Hugging Face organization, whose dataset metadata declares CC-BY-3.0. IBM’s publication page confirms this is the 30,497 crowd-sourced argument dataset released as part of the work. The card body licensing paragraph instead refers to “Copyright IBM 2014,” Wikipedia, and CC-BY-SA 3.0, which does not fit this dataset and conflicts with the frontmatter; preserve this discrepancy and do not substitute CC-BY-SA-3.0. - -- [Official IBM Research dataset page metadata declares cc-by-3.0.](https://huggingface.co/datasets/ibm-research/argument_quality_ranking_30k) -- [IBM Research publication identifies the dataset as 30,497 crowd-sourced arguments and says it was released as part of the work; it provides no alternate license terms.](https://research.ibm.com/publications/a-large-scale-dataset-for-argument-quality-ranking-construction-and-analysis) -- [Dataset card body contains conflicting legacy-looking terms (Copyright IBM 2014, Wikipedia, and CC-BY-SA 3.0) while frontmatter says CC-BY-3.0.](https://huggingface.co/datasets/ibm-research/argument_quality_ranking_30k/blob/590726b3765b1b90c5e53a17e3b1f77d92d3aa8a/README.md) - - -## jhu-clsp/FollowIR-train - -Source: [https://huggingface.co/datasets/jhu-clsp/FollowIR-train](https://huggingface.co/datasets/jhu-clsp/FollowIR-train) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -The live dataset card declares Apache-2.0. It says query/instruction pairs come from TREC track narratives and synthetic documents were generated with GPT-3.5-Turbo-1106 then filtered with Mistral; the card wording about “filtered instructions only” is inconsistent with its schema, which includes a document field. Preserve the TREC source rights caveat; no standalone TREC license was established here. - -- [Live card declares Apache-2.0; describes TREC query/narrative provenance and synthetic GPT-generated documents, and lists query, instruction, and document fields.](https://huggingface.co/datasets/jhu-clsp/FollowIR-train/raw/main/README.md) -- [FollowIR paper says the benchmark repurposes detailed instructions/narratives from three TREC shared-task collections.](https://arxiv.org/abs/2403.15246) - - -## lighteval/bbq_helm - -Source: [https://huggingface.co/datasets/lighteval/bbq_helm](https://huggingface.co/datasets/lighteval/bbq_helm) - -Recorded terms: **cc-by-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/lighteval/bbq_helm/blob/10937569dd76b6f33b2ff3572382ed87458e7856/README.md) - - -## lighteval/piqa - -Source: [https://huggingface.co/datasets/lighteval/piqa](https://huggingface.co/datasets/lighteval/piqa) - -Recorded terms: **afl-3.0**. Status: `verified`. - -Original PIQA project page explicitly identifies Academic Free License v3.0. - -- [Original project page’s License section identifies Academic Free License (AFL) v. 3.0.](https://yonatanbisk.com/piqa/) - - -## mandarjoshi/trivia_qa - -Source: [https://huggingface.co/datasets/mandarjoshi/trivia_qa](https://huggingface.co/datasets/mandarjoshi/trivia_qa) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -Project declares Apache-2.0 for dataset, but HF card warns that included questions/documents are not owned by UW; third-party source documents may carry separate rights. Treat as qualified, not blanket clearance. - -- [Official TriviaQA repository README states Apache 2.0 applies to code and data.](https://github.com/mandarjoshi90/triviaqa) -- [HF card identifies official project and explicitly states University of Washington does not own copyright of questions and documents included in TriviaQA.](https://huggingface.co/datasets/mandarjoshi/trivia_qa/raw/main/README.md) - - -## microsoft/ms_marco - -Source: [https://huggingface.co/datasets/microsoft/ms_marco](https://huggingface.co/datasets/microsoft/ms_marco) - -Recorded terms: **Microsoft Research dataset terms: non-commercial research only; no license or other IP rights granted**. Status: `qualified`. - -Microsoft’s official MS MARCO terms restrict use to non-commercial research and expressly state distribution is without extending any license or other intellectual-property rights. The HF mirror has no metadata license. - -- [Terms and Conditions: MS MARCO/ORCAS are “intended for non-commercial research purposes only” and made available “without extending any license or other intellectual property rights”; use binds users to those terms.](https://github.com/microsoft/msmarco/blob/master/README.md) -- [Official Microsoft Research page identifies MS MARCO and links to Access for MS MARCO.](https://www.microsoft.com/en-us/research/project/deep-learning-machine-reading-comprehension/tools/) - - -## miracl/miracl - -Source: [https://huggingface.co/datasets/miracl/miracl](https://huggingface.co/datasets/miracl/miracl) - -Recorded terms: **apache-2.0 (topics/qrels); Apache-2.0 declared for corpus package, with Wikipedia source-text terms (CC BY-SA 3.0/GFDL historical dump terms)**. Status: `qualified`. - -The retained source is topics/qrels, whose card declares Apache-2.0. Its sibling miracl/miracl-corpus declares Apache-2.0 but explicitly says documents are passages from Wikipedia dumps. MIRACL corpus was produced from older Wikipedia dumps, when Wikipedia text was under CC BY-SA 3.0 and GFDL; corpus/package Apache metadata must not be treated as overriding original text licenses. - -- [Topics/qrels repository card declares Apache-2.0 and says collection is separate.](https://huggingface.co/datasets/miracl/miracl/blob/main/README.md) -- [Corpus card declares Apache-2.0 and says each language corpus is prepared from a Wikipedia dump.](https://huggingface.co/datasets/miracl/miracl-corpus/blob/main/README.md) -- [Wikimedia terms state text is licensed under CC BY-SA and GFDL; historical dump date is relevant to version. MIRACL used older Wikipedia dumps, so retained Wikipedia text remains subject to source terms.](https://foundation.wikimedia.org/wiki/Terms_of_Use?useformat=mobile) - - -## mteb/Core17InstructionRetrieval - -Source: [https://huggingface.co/datasets/mteb/Core17InstructionRetrieval](https://huggingface.co/datasets/mteb/Core17InstructionRetrieval) - -Recorded terms: **mit**. Status: `qualified`. - -The MTEB task card declares MIT, and its source dataset card identifies the data as Core17 documents/queries. Core17 is based on the New York Times Annotated Corpus (LDC2008T19), not the TREC Washington Post corpus (that is Core18). Treat the MIT card declaration as applying to this release/annotations; underlying NYT articles remain governed by their corpus access terms. - -- [Task card frontmatter declares MIT and names jhu-clsp/core17-instructions-mteb as source.](https://huggingface.co/datasets/mteb/Core17InstructionRetrieval/raw/main/README.md) -- [Source card states its corpus config consists of original corpus documents and does not declare a separate license.](https://huggingface.co/datasets/jhu-clsp/core17-instructions-mteb/raw/main/README.md) -- [Official TREC 2017 Common Core page identifies its corpus as LDC2008T19.](https://pages.nist.gov/trec-browser/trec26/core/data/) -- [LDC catalog entry identifies LDC2008T19 as the New York Times Annotated Corpus.](https://catalog.ldc.upenn.edu/LDC2008T19) - - -## mteb/News21InstructionRetrieval - -Source: [https://huggingface.co/datasets/mteb/News21InstructionRetrieval](https://huggingface.co/datasets/mteb/News21InstructionRetrieval) - -Recorded terms: **MIT for the benchmark/derived release metadata; Washington Post corpus remains subject to its source agreement**. Status: `qualified`. - -The MTEB task card declares MIT, while its source_datasets field points to the JHU News21 instructions dataset. That source derives documents from the TREC Washington Post corpus; MIT on the task package does not relicense those article texts. NIST describes access by signed organizational and individual agreement. Keep release annotations and underlying WP corpus terms separate. - -- [MTEB card declares MIT and identifies jhu-clsp/news21-instructions-mteb as source dataset.](https://huggingface.co/datasets/mteb/News21InstructionRetrieval) -- [JHU source card describes reformatted queries, corpus, instructions and qrels; it does not include a data-specific license clarification in its body.](https://huggingface.co/datasets/jhu-clsp/news21-instructions-mteb) -- [NIST identifies the TREC Washington Post corpus as 728,626 WP news articles/blog posts and requires signed organizational and individual agreements for access.](https://trec.nist.gov/data/wapost/) - - -## mteb/Robust04InstructionRetrieval - -Source: [https://huggingface.co/datasets/mteb/Robust04InstructionRetrieval](https://huggingface.co/datasets/mteb/Robust04InstructionRetrieval) - -Recorded terms: **MIT (instruction/query additions); Robust04 underlying corpus is restricted research access**. Status: `qualified`. - -The HF card declares MIT, but corpus is Robust04/TREC Disks 4 and 5. NIST says these collections are copyrighted and should be used for research; access requires organizational/individual agreements. Thus MIT applies at most to this derived instruction package, not source articles. - -- [HF card declares MIT for derived instruction retrieval dataset.](https://huggingface.co/datasets/mteb/Robust04InstructionRetrieval/blob/0495a5cb69aa8fa5bca81bd56ac248911c19eeb6/README.md) -- [NIST says collections are copyrighted and commercially valuable and must be used for research purposes.](https://trec.nist.gov/data/docs_eng.html) -- [Robust04 requires TREC disks 4/5 from NIST and agreements; documents are copyrighted and corpus is for research use only.](https://ir-datasets.com/trec-robust04.html) - - -## nightingal3/fig-qa - -Source: [https://huggingface.co/datasets/nightingal3/fig-qa](https://huggingface.co/datasets/nightingal3/fig-qa) - -Recorded terms: **mit**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: - mit.](https://huggingface.co/datasets/nightingal3/fig-qa/raw/main/README.md) - - -## nomic-ai/nomic-embed-unsupervised-data - -Source: [https://huggingface.co/datasets/nomic-ai/nomic-embed-unsupervised-data](https://huggingface.co/datasets/nomic-ai/nomic-embed-unsupervised-data) - -Recorded terms: **unknown (source attribution and terms for codesearch subset unverified)**. Status: `unresolved`. - -Local source inventory confirms use of only `codesearch` and `synthetic_relevance__codesearch`. Nomic contrastors training config maps the `codesearch` label to an internal codesearch_full bucket but does not identify its source dataset or provide component license terms. Similarity of the name to CodeSearchNet is insufficient to assert provenance, and CoIR benchmark terms do not automatically govern this Nomic copy. Upstream attribution/license remains unresolved. - -- [Training data config maps `name: codesearch` to `s3://contrastive-index-filtered/codesearch_full/...`; it does not name an upstream dataset or license.](https://github.com/nomic-ai/contrastors/blob/main/src/contrastors/configs/data/contrastive_pretrain.yaml) -- [Card lists a codesearch data split/config and general examples from multiple source types but no source/license attribution specific to the codesearch subset.](https://huggingface.co/datasets/nomic-ai/nomic-embed-unsupervised-data) -- [Data Access section describes Nomic-hosted data access, not a license grant for the subset or a source-level rights statement.](https://github.com/nomic-ai/contrastors) - - -## nvidia/Aegis-AI-Content-Safety-Dataset-2.0 - -Source: [https://huggingface.co/datasets/nvidia/Aegis-AI-Content-Safety-Dataset-2.0](https://huggingface.co/datasets/nvidia/Aegis-AI-Content-Safety-Dataset-2.0) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['cc-by-4.0']; the card body/context is preserved in 25-README.md.](https://huggingface.co/datasets/nvidia/Aegis-AI-Content-Safety-Dataset-2.0/blob/d86bb8bedff51d25ac834ab7838f1cc61acb7a2c/README.md) - - -## nvidia/CantTalkAboutThis-Topic-Control-Dataset - -Source: [https://huggingface.co/datasets/nvidia/CantTalkAboutThis-Topic-Control-Dataset](https://huggingface.co/datasets/nvidia/CantTalkAboutThis-Topic-Control-Dataset) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: cc-by-4.0.](https://huggingface.co/datasets/nvidia/CantTalkAboutThis-Topic-Control-Dataset/raw/main/README.md) - - -## nvidia/HelpSteer - -Source: [https://huggingface.co/datasets/nvidia/HelpSteer](https://huggingface.co/datasets/nvidia/HelpSteer) - -Recorded terms: **cc-by-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/nvidia/HelpSteer/blob/3ca5d59c1bc1080af195b4254e7407db60b6f450/README.md) - - -## nvidia/HelpSteer2 - -Source: [https://huggingface.co/datasets/nvidia/HelpSteer2](https://huggingface.co/datasets/nvidia/HelpSteer2) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['cc-by-4.0']; the card body/context is preserved in 26-README.md.](https://huggingface.co/datasets/nvidia/HelpSteer2/blob/990b2711a36180dd19d9c94b8627844866f8982a/README.md) - - -## nvidia/HelpSteer3 - -Source: [https://huggingface.co/datasets/nvidia/HelpSteer3](https://huggingface.co/datasets/nvidia/HelpSteer3) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: cc-by-4.0.](https://huggingface.co/datasets/nvidia/HelpSteer3/raw/main/README.md) - - -## nyu-mll/blimp - -Source: [https://huggingface.co/datasets/nyu-mll/blimp](https://huggingface.co/datasets/nyu-mll/blimp) - -Recorded terms: **cc-by-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/nyu-mll/blimp/blob/877fba0801ffb7cbd8c39c1ff314a46f053f6036/README.md) - - -## nyu-mll/multi_nli - -Source: [https://huggingface.co/datasets/nyu-mll/multi_nli](https://huggingface.co/datasets/nyu-mll/multi_nli) - -Recorded terms: **cc-by-3.0; cc-by-sa-3.0; additional source-specific terms**. Status: `qualified`. - -Card describes multiple underlying Open American National Corpus and fiction licenses; metadata has several identifiers and an “other” value. Treat the included text as carrying source-specific terms, not as one blanket license. - -- [Card details Open American National Corpus terms and heterogeneous fiction permissions: CC-BY-SA 3.0, CC-BY 3.0, and public domain for some works; it also says licenses may differ by jurisdiction.](https://huggingface.co/datasets/nyu-mll/multi_nli/blob/da70db2af9d09693783c3320c4249840212ee221/README.md) - - -## open-index/open-arxiv - -Source: [https://huggingface.co/datasets/open-index/open-arxiv](https://huggingface.co/datasets/open-index/open-arxiv) - -Recorded terms: **cc0-1.0 (metadata only; per-paper terms vary)**. Status: `qualified`. - -The card says it contains arXiv metadata only and is derived from Cornell’s Kaggle arXiv metadata dataset under CC0 1.0. It explicitly says papers have individual licenses in the record `license` field; CC0 does not extend to paper full text or source files. The used metadata rows still include abstracts, so retain the dataset’s stated metadata CC0 scope and paper-level field rather than treating the articles as CC0. - -- [Card states metadata-only, CC0 1.0, and that individual papers carry their own terms in the license field.](https://huggingface.co/datasets/open-index/open-arxiv/raw/main/README.md) -- [Original Cornell Kaggle arXiv dataset card says CC0 1.0 applies to metadata and points to arXiv license information for individual papers.](https://www.kaggle.com/datasets/Cornell-University/arxiv) -- [arXiv explains articles retain their own author-selected license/copyright terms.](https://arxiv.org/help/license) - - -## openai/gsm8k - -Source: [https://huggingface.co/datasets/openai/gsm8k](https://huggingface.co/datasets/openai/gsm8k) - -Recorded terms: **mit**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license mit; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/openai/gsm8k/blob/740312add88f781978c0658806c59bc2815b9866/README.md) - - -## qiaojin/PubMedQA - -Source: [https://huggingface.co/datasets/qiaojin/PubMedQA](https://huggingface.co/datasets/qiaojin/PubMedQA) - -Recorded terms: **MIT (HF card declaration); PubMed abstract text rights remain source-specific**. Status: `qualified`. - -HF card declares MIT, and its data includes PubMed abstracts. NLM says it does not claim copyright on abstracts; publishers/authors may hold copyright. Do not treat MIT as a blanket grant for abstract text. - -- [Card metadata declares MIT; dataset card describes PubMedQA records with PubMed-derived contexts/abstracts.](https://huggingface.co/datasets/qiaojin/PubMedQA/blob/main/README.md) -- [NLM says it does not claim copyright on PubMed abstracts; publishers or authors may.](https://www.ncbi.nlm.nih.gov/home/about/policies/) - - -## rajpurkar/squad - -Source: [https://huggingface.co/datasets/rajpurkar/squad](https://huggingface.co/datasets/rajpurkar/squad) - -Recorded terms: **cc-by-sa-4.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: cc-by-sa-4.0.](https://huggingface.co/datasets/rajpurkar/squad/raw/main/README.md) - - -## rajpurkar/squad_v2 - -Source: [https://huggingface.co/datasets/rajpurkar/squad_v2](https://huggingface.co/datasets/rajpurkar/squad_v2) - -Recorded terms: **cc-by-sa-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-sa-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/rajpurkar/squad_v2/blob/3ffb306f725f7d2ce8394bc1873b24868140c412/README.md) - - -## sentence-transformers/agnews - -Source: [https://huggingface.co/datasets/sentence-transformers/agnews](https://huggingface.co/datasets/sentence-transformers/agnews) - -Recorded terms: **unknown**. Status: `unresolved`. - -Reviewed original AG News corpus homepage, HF card, and Sentence Transformers embedding-training-data notice. The distributor expressly disclaims granting dataset licenses; no AG News content license was found. - -- [Original AG News corpus page provides no license grant.](http://groups.di.unipi.it/~gulli/AG_corpus_of_news_articles.html) -- [Distributor says it does not claim users have a license; responsibility to determine permission remains with users.](https://huggingface.co/datasets/sentence-transformers/embedding-training-data) - - -## sentence-transformers/all-nli - -Source: [https://huggingface.co/datasets/sentence-transformers/all-nli](https://huggingface.co/datasets/sentence-transformers/all-nli) - -Recorded terms: **cc-by-sa-4.0; cc-by-sa-3.0; cc-by-3.0; public-domain; OANC terms**. Status: `qualified`. - -Composite uses SNLI and MultiNLI. SNLI is CC BY-SA 4.0. The MultiNLI paper Distribution section describes mixed upstream terms: most sources are OANC terms; fiction Seven Swords is CC BY-SA 3.0; Living History and Password Incorrect are CC BY 3.0; other text is US public domain. Do not summarize the corpus with one blanket license; retain source-specific terms. - -- [Card says AllNLI is a concatenation of SNLI and MultiNLI.](https://huggingface.co/datasets/sentence-transformers/all-nli/raw/main/README.md) -- [Original SNLI project states corpus is CC BY-SA 4.0.](https://nlp.stanford.edu/projects/snli/) -- [Original MultiNLI paper, Distribution section, describes different terms for constituent genres: OANC terms, CC BY-SA 3.0, CC BY 3.0, and US public domain.](https://cims.nyu.edu/~sbowman/multinli/paper.pdf) - - -## sentence-transformers/coco-captions - -Source: [https://huggingface.co/datasets/sentence-transformers/coco-captions](https://huggingface.co/datasets/sentence-transformers/coco-captions) - -Recorded terms: **CC BY 4.0 (COCO annotations/captions); image rights remain with Flickr contributors**. Status: `qualified`. - -This HF dataset contains caption annotation text. Official COCO terms license annotations under CC BY 4.0; they explicitly say COCO does not own image copyrights and image use must follow Flickr terms. Do not apply CC BY 4.0 to COCO images. - -- [Terms of Use > Annotations & Website: annotations licensed CC BY 4.0. Terms of Use > Images: COCO Consortium does not own image copyrights; image use must abide by Flickr Terms of Use.](https://github.com/cocodataset/cocodataset.github.io/blob/master/dataset/termsofuse.htm) -- [HF card describes the caption dataset but has no independent license declaration.](https://huggingface.co/datasets/sentence-transformers/coco-captions) - - -## sentence-transformers/gooaq - -Source: [https://huggingface.co/datasets/sentence-transformers/gooaq](https://huggingface.co/datasets/sentence-transformers/gooaq) - -Recorded terms: **non-commercial restriction (source README); Apache-2.0 LICENSE file scope unclear**. Status: `qualified`. - -AllenAI GooAQ source README says it should not be used for commercial purposes and points to LICENSE. The LICENSE is Apache-2.0, but the README restriction means its scope over dataset content needs careful reading; do not treat Apache code license alone as commercial clearance. HF embedding-training-data explicitly says it does not grant or vouch for dataset licenses. - -- [Source README explicitly states: “This dataset should not be used for any commercial purposes” and links its LICENSE.](https://github.com/allenai/gooaq) -- [LICENSE file is Apache License 2.0; because the README separately restricts dataset commercial use, scope is qualified.](https://github.com/allenai/gooaq/blob/main/LICENSE) -- [The distributor says it does not claim users have a license and users are responsible for determining permission.](https://huggingface.co/datasets/sentence-transformers/embedding-training-data) - - -## sentence-transformers/natural-questions - -Source: [https://huggingface.co/datasets/sentence-transformers/natural-questions](https://huggingface.co/datasets/sentence-transformers/natural-questions) - -Recorded terms: **cc-by-sa-3.0**. Status: `verified`. - -Verified at the official Google dataset distribution. NQ includes rendered Wikipedia page text; source page identifies the dataset license as CC BY-SA 3.0. - -- [Official Google Natural Questions download page explicitly says Natural Questions is released under Creative Commons Share-Alike 3.0.](https://ai.google.com/research/NaturalQuestions/download) -- [HF card points to Natural Questions as upstream source.](https://huggingface.co/datasets/sentence-transformers/natural-questions/raw/main/README.md) - - -## sentence-transformers/trivia-qa - -Source: [https://huggingface.co/datasets/sentence-transformers/trivia-qa](https://huggingface.co/datasets/sentence-transformers/trivia-qa) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -The original TriviaQA repository identifies Apache-2.0, but the HF copy contains evidence passages drawn from third-party web pages; that repository license may not grant rights to the underlying evidence text. - -- [Original TriviaQA repository’s LICENSE is Apache License 2.0; the HF card includes web evidence text and does not establish that third-party evidence pages share that license.](https://github.com/mandarjoshi90/triviaqa/blob/master/LICENSE) - - -## slinusc/PubMedAbstractsSubset - -Source: [https://huggingface.co/datasets/slinusc/PubMedAbstractsSubset](https://huggingface.co/datasets/slinusc/PubMedAbstractsSubset) - -Recorded terms: **CC0-1.0 for released metadata declaration; abstract text rights remain source-specific**. Status: `qualified`. - -The card declares CC0-1.0, but NLM states it does not claim copyright on PubMed abstracts and publishers/authors may hold copyright, and advises users that it provides no legal advice. Therefore CC0 cannot be treated as a blanket grant for abstract text. This caveat also applies to hotchpotch/pubmed-abstract-ir-simulated-search-queries derived from this source. - -- [Card declares CC0-1.0 and says records include title, abstract, and PMID from NLM PubMed metadata.](https://huggingface.co/datasets/slinusc/PubMedAbstractsSubset/blob/111f2d4a15b4335a4ddf6844264e6caf2d0a614a/README.md) -- [NCBI/NLM says NLM does not claim copyright on PubMed abstracts; journal publishers or authors may, and NLM provides no legal advice about distribution.](https://www.ncbi.nlm.nih.gov/home/about/policies/) - - -## stanfordnlp/snli - -Source: [https://huggingface.co/datasets/stanfordnlp/snli](https://huggingface.co/datasets/stanfordnlp/snli) - -Recorded terms: **cc-by-sa-4.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: - cc-by-sa-4.0.](https://huggingface.co/datasets/stanfordnlp/snli/raw/main/README.md) - - -## tasksource/Boardgame-QA - -Source: [https://huggingface.co/datasets/tasksource/Boardgame-QA](https://huggingface.co/datasets/tasksource/Boardgame-QA) - -Recorded terms: **cc-by-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/tasksource/Boardgame-QA/blob/78e38c3c8df3b4f6de7ae8bd1fc6a8bd1f31be56/README.md) - - -## tasksource/FOL-nli - -Source: [https://huggingface.co/datasets/tasksource/FOL-nli](https://huggingface.co/datasets/tasksource/FOL-nli) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['apache-2.0']; the card body/context is preserved in 32-README.md.](https://huggingface.co/datasets/tasksource/FOL-nli/blob/6b8a2ec01b226ed871fbde4677a31146ac5f370c/README.md) - - -## tasksource/PLANE-ood - -Source: [https://huggingface.co/datasets/tasksource/PLANE-ood](https://huggingface.co/datasets/tasksource/PLANE-ood) - -Recorded terms: **cc-by-2.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: cc-by-2.0.](https://huggingface.co/datasets/tasksource/PLANE-ood/raw/main/README.md) - - -## tasksource/SDOH-NLI - -Source: [https://huggingface.co/datasets/tasksource/SDOH-NLI](https://huggingface.co/datasets/tasksource/SDOH-NLI) - -Recorded terms: **cc-by-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/tasksource/SDOH-NLI/blob/75d4cc5f67ecb3f3c44cf5a3fadc3c9579f2f148/README.md) - - -## tasksource/SpaceNLI - -Source: [https://huggingface.co/datasets/tasksource/SpaceNLI](https://huggingface.co/datasets/tasksource/SpaceNLI) - -Recorded terms: **mit**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['mit']; the card body/context is preserved in 33-README.md.](https://huggingface.co/datasets/tasksource/SpaceNLI/blob/4755020cda14b4ca1f7d464430aa72b3acb8eadb/README.md) - - -## tasksource/arct - -Source: [https://huggingface.co/datasets/tasksource/arct](https://huggingface.co/datasets/tasksource/arct) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: apache-2.0.](https://huggingface.co/datasets/tasksource/arct/raw/main/README.md) - - -## tasksource/babi_nli - -Source: [https://huggingface.co/datasets/tasksource/babi_nli](https://huggingface.co/datasets/tasksource/babi_nli) - -Recorded terms: **bsd**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license bsd; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/tasksource/babi_nli/blob/80b689dc668b50f3640fc5368fd1893d385cb4e1/README.md) - - -## tasksource/commonsense_qa_2.0 - -Source: [https://huggingface.co/datasets/tasksource/commonsense_qa_2.0](https://huggingface.co/datasets/tasksource/commonsense_qa_2.0) - -Recorded terms: **cc-by-4.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['cc-by-4.0']; the card body/context is preserved in 34-README.md.](https://huggingface.co/datasets/tasksource/commonsense_qa_2.0/blob/23ff83b18ed76882f3af1a403a7c464b72efcd86/README.md) - - -## tasksource/corr2cause - -Source: [https://huggingface.co/datasets/tasksource/corr2cause](https://huggingface.co/datasets/tasksource/corr2cause) - -Recorded terms: **mit**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: mit.](https://huggingface.co/datasets/tasksource/corr2cause/raw/main/README.md) - - -## tasksource/defeasible-nli - -Source: [https://huggingface.co/datasets/tasksource/defeasible-nli](https://huggingface.co/datasets/tasksource/defeasible-nli) - -Recorded terms: **apache-2.0; cc-by-sa-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license apache-2.0 and cc-by-sa-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/tasksource/defeasible-nli/blob/7c4a57df9d8de5c36d4e9caa977907b5e8469c4f/README.md) - - -## tasksource/esci - -Source: [https://huggingface.co/datasets/tasksource/esci](https://huggingface.co/datasets/tasksource/esci) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['apache-2.0']; the card body/context is preserved in 35-README.md.](https://huggingface.co/datasets/tasksource/esci/blob/8113b17a5d4099e20243282c926f1bc1a08a4d13/README.md) - - -## tasksource/feasibilityQA - -Source: [https://huggingface.co/datasets/tasksource/feasibilityQA](https://huggingface.co/datasets/tasksource/feasibilityQA) - -Recorded terms: **mit**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: mit.](https://huggingface.co/datasets/tasksource/feasibilityQA/raw/main/README.md) - - -## tasksource/hans - -Source: [https://huggingface.co/datasets/tasksource/hans](https://huggingface.co/datasets/tasksource/hans) - -Recorded terms: **mit**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license mit; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/tasksource/hans/blob/8f107047d34fdfb57abfde38ec5196d63f7dee33/README.md) - - -## tasksource/lexcomp-nc-relation - -Source: [https://huggingface.co/datasets/tasksource/lexcomp-nc-relation](https://huggingface.co/datasets/tasksource/lexcomp-nc-relation) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['apache-2.0']; the card body/context is preserved in 36-README.md.](https://huggingface.co/datasets/tasksource/lexcomp-nc-relation/blob/4f7fa5d487d71ac1543bc7849993b5dfc2e361ac/README.md) - - -## tasksource/logical-entailment - -Source: [https://huggingface.co/datasets/tasksource/logical-entailment](https://huggingface.co/datasets/tasksource/logical-entailment) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: apache-2.0.](https://huggingface.co/datasets/tasksource/logical-entailment/raw/main/README.md) - - -## tasksource/lonli - -Source: [https://huggingface.co/datasets/tasksource/lonli](https://huggingface.co/datasets/tasksource/lonli) - -Recorded terms: **mit**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license mit; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/tasksource/lonli/blob/f17e55b8f60da2f32720c1a16e4b6149145c662f/README.md) - - -## tasksource/miqa - -Source: [https://huggingface.co/datasets/tasksource/miqa](https://huggingface.co/datasets/tasksource/miqa) - -Recorded terms: **cc-by-4.0 (HF card declaration)**. Status: `qualified`. - -Fetched source snapshot frontmatter says only `license: cc-by-4.0`; imported metadata also included Apache-2.0 without support from this card. Underlying MIQA sources may carry additional terms; the card does not establish an Apache-2.0 grant. - -- [Actual fetched README is just frontmatter declaring CC-BY-4.0.](https://huggingface.co/datasets/tasksource/miqa/blob/b0b4a19860eea1740f04a7836612ec239e08e838/README.md) - - -## tasksource/monotonicity-entailment - -Source: [https://huggingface.co/datasets/tasksource/monotonicity-entailment](https://huggingface.co/datasets/tasksource/monotonicity-entailment) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: apache-2.0.](https://huggingface.co/datasets/tasksource/monotonicity-entailment/raw/main/README.md) - - -## tasksource/mtop - -Source: [https://huggingface.co/datasets/tasksource/mtop](https://huggingface.co/datasets/tasksource/mtop) - -Recorded terms: **cc-by-sa-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-sa-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/tasksource/mtop/blob/9fa95557ba2809ad05943880eee5231b4f2463fb/README.md) - - -## tasksource/naturallogic - -Source: [https://huggingface.co/datasets/tasksource/naturallogic](https://huggingface.co/datasets/tasksource/naturallogic) - -Recorded terms: **apache-2.0 (HF card declaration)**. Status: `qualified`. - -Fetched source snapshot frontmatter says Apache-2.0 only. The second CC-BY-SA-4.0 identifier in imported metadata is unsupported by this card; the linked original project should be checked for source-level terms. - -- [Actual fetched card frontmatter declares Apache-2.0 only.](https://huggingface.co/datasets/tasksource/naturallogic/blob/746f83a1f110ee7cd4ca4267f28a9ef044fb8d4b/README.md) - - -## tasksource/nlsat - -Source: [https://huggingface.co/datasets/tasksource/nlsat](https://huggingface.co/datasets/tasksource/nlsat) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: apache-2.0.](https://huggingface.co/datasets/tasksource/nlsat/raw/main/README.md) - - -## tasksource/patent-phrase-similarity - -Source: [https://huggingface.co/datasets/tasksource/patent-phrase-similarity](https://huggingface.co/datasets/tasksource/patent-phrase-similarity) - -Recorded terms: **cc-by-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/tasksource/patent-phrase-similarity/blob/ff1868796e44ac9ec02e870b447fde6c5bf2d6e4/README.md) - - -## tasksource/robustLR - -Source: [https://huggingface.co/datasets/tasksource/robustLR](https://huggingface.co/datasets/tasksource/robustLR) - -Recorded terms: **mit**. Status: `verified`. - -HF card itself omits a license, but original INK-USC RobustLR repository has a MIT License naming INK Lab @ USC and 2022. Snapshotted upstream LICENSE in evidence/batch-1/deep-2.txt. - -- [Original LICENSE begins “MIT License” and “Copyright (c) 2022 INK Lab @ USC”.](https://github.com/INK-USC/RobustLR/blob/main/LICENSE) - - -## tasksource/ruletaker - -Source: [https://huggingface.co/datasets/tasksource/ruletaker](https://huggingface.co/datasets/tasksource/ruletaker) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: apache-2.0.](https://huggingface.co/datasets/tasksource/ruletaker/raw/main/README.md) - - -## tasksource/scicite - -Source: [https://huggingface.co/datasets/tasksource/scicite](https://huggingface.co/datasets/tasksource/scicite) - -Recorded terms: **unknown (original dataset license not located)**. Status: `unresolved`. - -The live tasksource card labels the license unknown. AllenAI SciCite’s original README explains how to obtain the dataset but does not give dataset terms; the repository’s code license does not grant rights to the citation contexts or source papers. - -- [Live card frontmatter records license: unknown.](https://huggingface.co/datasets/tasksource/scicite) -- [Original dataset README describes the SciCite data and download but has no dataset license/redistribution grant.](https://github.com/allenai/scicite) - - -## tasksource/scone - -Source: [https://huggingface.co/datasets/tasksource/scone](https://huggingface.co/datasets/tasksource/scone) - -Recorded terms: **cc0-1.0**. Status: `qualified`. - -License is explicitly declared in the cited source dataset card/metadata. This audit did not establish a separate blanket license for every upstream third-party component or source work; check the linked original dataset terms before reuse. - -- [Fetched and reviewed source card at pinned revision. Its frontmatter license field declares ['cc0-1.0']; the card body/context is preserved in 40-README.md.](https://huggingface.co/datasets/tasksource/scone/blob/505ed270f9e9865c6cfab1e222d804fedd837127/README.md) - - -## tasksource/spartqa-mchoice - -Source: [https://huggingface.co/datasets/tasksource/spartqa-mchoice](https://huggingface.co/datasets/tasksource/spartqa-mchoice) - -Recorded terms: **mit**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: mit.](https://huggingface.co/datasets/tasksource/spartqa-mchoice/raw/main/README.md) - - -## tasksource/stepgame - -Source: [https://huggingface.co/datasets/tasksource/stepgame](https://huggingface.co/datasets/tasksource/stepgame) - -Recorded terms: **mit**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license mit; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/tasksource/stepgame/blob/3ba235d4a7056cf8b062a57aa489f7bc14bff229/README.md) - - -## tasksource/strategy-qa - -Source: [https://huggingface.co/datasets/tasksource/strategy-qa](https://huggingface.co/datasets/tasksource/strategy-qa) - -Recorded terms: **mit**. Status: `verified`. - -Pinned snapshot URL returned 404, but the original official StrategyQA repository by the authors contains an MIT LICENSE. This is upstream evidence, not the failed snapshot. - -- [Official authors’ repository LICENSE is MIT and names Elad Segal, 2021.](https://github.com/eladsegal/strategyqa/blob/main/LICENSE) - - -## tasksource/temporal-nli - -Source: [https://huggingface.co/datasets/tasksource/temporal-nli](https://huggingface.co/datasets/tasksource/temporal-nli) - -Recorded terms: **apache-2.0**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: apache-2.0.](https://huggingface.co/datasets/tasksource/temporal-nli/raw/main/README.md) - - -## tasksource/tracie - -Source: [https://huggingface.co/datasets/tasksource/tracie](https://huggingface.co/datasets/tasksource/tracie) - -Recorded terms: **apache-2.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license apache-2.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/tasksource/tracie/blob/4967d8aaa06a4ea97a88bf77071d66be0140bc45/README.md) - - -## tasksource/winowhy - -Source: [https://huggingface.co/datasets/tasksource/winowhy](https://huggingface.co/datasets/tasksource/winowhy) - -Recorded terms: **mit**. Status: `qualified`. - -Original WinoWhy repository LICENSE is MIT. HF snapshot only has MIT frontmatter and the repository includes added generated explanations; verify that source materials permit the intended downstream use. - -- [Original repository LICENSE is MIT and names HKUST-KnowComp, 2020.](https://github.com/HKUST-KnowComp/WinoWhy/blob/master/LICENSE) - - -## tau/commonsense_qa - -Source: [https://huggingface.co/datasets/tau/commonsense_qa](https://huggingface.co/datasets/tau/commonsense_qa) - -Recorded terms: **mit**. Status: `qualified`. - -License recorded from source dataset repository metadata; upstream provenance and any embedded third-party content require dataset-specific scope review. - -- [Live dataset README frontmatter declares license: - mit.](https://huggingface.co/datasets/tau/commonsense_qa/raw/main/README.md) - - -## ucberkeley-dlab/measuring-hate-speech - -Source: [https://huggingface.co/datasets/ucberkeley-dlab/measuring-hate-speech](https://huggingface.co/datasets/ucberkeley-dlab/measuring-hate-speech) - -Recorded terms: **cc-by-4.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license cc-by-4.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/ucberkeley-dlab/measuring-hate-speech/blob/5468f6e118396646b02a2f691e771f6b6d9502ea/README.md) - - -## ucirvine/sms_spam - -Source: [https://huggingface.co/datasets/ucirvine/sms_spam](https://huggingface.co/datasets/ucirvine/sms_spam) - -Recorded terms: **cc-by-4.0**. Status: `verified`. - -Original UCI dataset page explicitly identifies the SMS Spam Collection as licensed under Creative Commons Attribution 4.0 International. - -- [Original UCI dataset page states the collection is available under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.](https://archive.ics.uci.edu/dataset/228/sms+spam+collection) - - -## universal-dependencies/universal_dependencies - -Source: [https://huggingface.co/datasets/universal-dependencies/universal_dependencies](https://huggingface.co/datasets/universal-dependencies/universal_dependencies) - -Recorded terms: **cc-by-sa-4.0**. Status: `verified`. - -The active input config is ud_ewt (English-EWT). Its upstream treebank-specific license page and LICENSE.txt both specify CC BY-SA 4.0. Other UD treebanks carry separate terms; they are outside this active subset. - -- [Official English-EWT treebank page explicitly states License: CC BY-SA 4.0.](https://universaldependencies.org/treebanks/en_ewt/index.html) -- [Actual upstream English-EWT LICENSE.txt is the Creative Commons Attribution-ShareAlike 4.0 International license text.](https://github.com/UniversalDependencies/UD_English-EWT/blob/master/LICENSE.txt) - - -## zefang-liu/phishing-email-dataset - -Source: [https://huggingface.co/datasets/zefang-liu/phishing-email-dataset](https://huggingface.co/datasets/zefang-liu/phishing-email-dataset) - -Recorded terms: **lgpl-3.0**. Status: `verified`. - - - -- [Reviewed source card/repository states dataset license lgpl-3.0; retained card text is in the evidence snapshot.](https://huggingface.co/datasets/zefang-liu/phishing-email-dataset/blob/main/README.md)