Datasets:
text stringlengths 0 109k |
|---|
SOURCE_URL: https://aleph-alpha.com/terms-conditions/ |
FINAL_URL: https://aleph-alpha.com/ |
HTTP_STATUS: 200 |
ARTEFACT_TYPE: terms_of_service |
ACCESSED: 2026-08-04 |
---------------------------------------------------------------------- |
Aleph Alpha Blog Kontakt Intern De En Vertrauen. Verantwortung. Souveränität. 01 Vision Spezialisierte Sprachmodelle für ein souveränes Europa. In enger Partnerschaft mit unseren Kunden konzipieren, entwickeln und betreiben wir individualisierte, sichere Sprachmodelle (SLLMs) für reale, geschäftskritische Umgebungen un... |
SOURCE_URL: https://aleph-alpha.com/en/ |
FINAL_URL: https://aleph-alpha.com/en/ |
HTTP_STATUS: 200 |
ARTEFACT_TYPE: public_docs |
ACCESSED: 2026-08-04 |
---------------------------------------------------------------------- |
Aleph Alpha Blog Contact Internal De En Trust. Responsibility. Sovereignty. 01 Vision We create Specialized Large Language Models for a sovereign Europe. We design, build, and deploy SLLMs in close partnership with our customers, taking responsibility for their performance in real-world, mission-critical environments. ... |
SOURCE_URL: https://huggingface.co/Aleph-Alpha |
FINAL_URL: https://huggingface.co/Aleph-Alpha |
HTTP_STATUS: 200 |
ARTEFACT_TYPE: model_card |
ACCESSED: 2026-08-04 |
---------------------------------------------------------------------- |
Aleph-Alpha (Aleph Alpha) |
Hugging Face Models Datasets Spaces Buckets new Docs Enterprise Pricing Website Tasks HuggingChat Collections Languages Organizations Community Blog Posts Daily Papers Hardware Learn Discord Forum GitHub Solutions Team & Enterprise Hugging Face PRO Enterprise Support Inference Providers Inference Endpoints Storage Buc... |
llama-3_1-tfree-hat models: This model family replaces the Llama 3.1 tokenizer with our HAT architecture. The 8b-dpo model is tuned for helpfulness and reduced refusal in sensitive applications, while the larger 70b-sft model is trained on English/German for improved text compression and adaptability. |
TFree-HAT-Pretrained-7B-Base: This 7B model was pretrained from scratch in English & German and has a context length of 32,900 words. It shows strong proficiency in German and beats Llama 3.1 on many English benchmarks. |
We also published a SOTA German Dataset ( data , arXiv ), which can be used to enhance German LLM capabilities. |
Our future work is dedicated to advancing reasoning models, de-biasing frontier models, understanding the role of data in model training, comprehensive and realistic model evaluation, pushing the boundaries of small models, and advancing tokenizer-free architectures. We will continue to concentrate on creating transpa... |
Want to shape the future of sovereign AI? Work with us . |
Collections 4 Llama-3.1-70B Tfree HAT Aleph-Alpha/llama-3_1-70b-tfree-hat-sft 69B • Updated Sep 15, 2025 • 19 • 2 Tfree-HAT-7b-pretrained Tokenizer free models based on Hierarchical Autoregressive Transformer (https://arxiv.org/abs/2501.10322) trained from scratch. Aleph-Alpha/llama-tfree-hat-pretrained-7b-dpo 7B • Up... |
Recently updated pinned Sleeping Meta Eval Argilla ✍ Aleph-Alpha Sep 25, 2025 models 29 Sort: |
Recently updated Aleph-Alpha/tfree-hat-pretrained-7b-base 7B • Updated Oct 22, 2025 • 21 • 17 Aleph-Alpha/llama-tfree-hat-pretrained-7b-dpo 7B • Updated Oct 22, 2025 • 23 • 10 Aleph-Alpha/llama-3_1-8b-tfree-hat-dpo 7B • Updated Oct 22, 2025 • 30 • 16 Aleph-Alpha/llama-3_1-8b-tfree-hat-base 7B • Updated Oct 22, 2025 • ... |
models datasets 3 Sort: |
Recently updated Aleph-Alpha/Aleph-Alpha-GermanWeb Viewer • Updated Mar 31 • 1.43B • 245 • 23 Aleph-Alpha/MTBench-German Updated Apr 17, 2025 • 57 Aleph-Alpha/ticket-classification Updated Mar 14, 2024 • 25 System theme Company TOS Privacy About Careers Website Models Datasets Spaces Pricing Docs |
SOURCE_URL: https://openai.com/policies/terms-of-use/ |
FINAL_URL: https://openai.com/policies/terms-of-use/ |
HTTP_STATUS: 200 |
ARTEFACT_TYPE: terms_of_service |
ACCESSED: 2026-08-04 |
---------------------------------------------------------------------- |
Europe Terms of Use | OpenAI Skip to main content Research Products Business Developers Company Foundation (opens in a new window) Log in Try ChatGPT (opens in a new window) Research Products Business Developers Company Foundation (opens in a new window) Try ChatGPT (opens in a new window) Login OpenAI Select language ... |
SOURCE_URL: https://openai.com/policies/usage-policies/ |
FINAL_URL: https://openai.com/policies/usage-policies/ |
HTTP_STATUS: 200 |
ARTEFACT_TYPE: terms_of_service |
ACCESSED: 2026-08-04 |
---------------------------------------------------------------------- |
Usage policies | OpenAI Skip to main content Research Products Business Developers Company Foundation (opens in a new window) Log in Try ChatGPT (opens in a new window) Research Products Business Developers Company Foundation (opens in a new window) Try ChatGPT (opens in a new window) Login OpenAI Select language … Eff... |
SOURCE_URL: https://help.openai.com/en/articles/6783457-what-is-chatgpt |
FINAL_URL: https://help.openai.com/en/articles/6783457-what-is-chatgpt |
HTTP_STATUS: 200 |
ARTEFACT_TYPE: public_docs |
ACCESSED: 2026-08-04 |
---------------------------------------------------------------------- |
What is ChatGPT? | OpenAI Help Center |
All Collections ChatGPT What is ChatGPT? What is ChatGPT? Commonly asked questions about ChatGPT Updated: 6 days ago How much does it cost to use ChatGPT? ChatGPT is free to use, though we do have subscription plans available, available on our pricing page . How does ChatGPT work? ChatGPT is fine-tuned from GPT-3.5, a... |
SOURCE_URL: https://openai.com/index/understanding-the-source-of-what-we-see-and-hear-online/ |
FINAL_URL: https://openai.com/index/understanding-the-source-of-what-we-see-and-hear-online/ |
HTTP_STATUS: 200 |
ARTEFACT_TYPE: public_docs |
ACCESSED: 2026-08-04 |
---------------------------------------------------------------------- |
Understanding the source of what we see and hear online | OpenAI Skip to main content Research Products Business Developers Company Foundation (opens in a new window) Log in Try ChatGPT (opens in a new window) Research Products Business Developers Company Foundation (opens in a new window) Try ChatGPT (opens in a new w... |
SOURCE_URL: https://help.openai.com/en/articles/8912793-c2pa-in-dall-e-3 |
FINAL_URL: https://help.openai.com/en/articles/8912793-c2pa-in-dall-e-3 |
HTTP_STATUS: 200 |
ARTEFACT_TYPE: public_docs |
ACCESSED: 2026-08-04 |
---------------------------------------------------------------------- |
Provenance signals (Content Credentials, SynthID) in OpenAI-generated content | OpenAI Help Center |
All Collections Privacy and policies Policy FAQ Provenance signals (Content Credentials, SynthID) in OpenAI-generated content Provenance signals (Content Credentials, SynthID) in OpenAI-generated content Updated: 3 days ago OpenAI uses provenance signals such as SynthID watermarks and Content Credentials (C2PA) to hel... |
SOURCE_URL: https://openai.com/index/gpt-4o-system-card/ |
FINAL_URL: https://openai.com/index/gpt-4o-system-card/ |
HTTP_STATUS: 200 |
ARTEFACT_TYPE: model_card |
ACCESSED: 2026-08-04 |
---------------------------------------------------------------------- |
GPT-4o System Card | OpenAI Skip to main content Research Products Business Developers Company Foundation (opens in a new window) Log in Try ChatGPT (opens in a new window) Research Products Business Developers Company Foundation (opens in a new window) Try ChatGPT (opens in a new window) Login OpenAI August 8, 2024 Pu... |
Article AI |
Article AI |
Chatbot We evaluated the persuasiveness of GPT‑4o’s text and voice modalities. Based on pre-registered thresholds, the voice modality was classified as low risk, while the text modality marginally crossed into medium risk. For the text modality, we evaluated the persuasiveness of GPT‑4o‑generated articles and chatbots ... |
Optimization , EAAMO ’21, ACM, Oct. 2021. 16 S. Shahriar, S. Allana, S. M. Hazratifard, and R. Dara, “A survey of privacy risks and mitigation strategies in the artificial intelligence life cycle,” IEEE Access, vol. 11, pp. 61829– 61854, 2023. 17 OpenAI, “Moderation overview,” 2024. 18 A. Tamkin, M. Brundage, J. Clark,... |
science and public health capabilities,” 2024. https://www2.deloitte.com/us/en/pages/about-deloitte/articles/press-releases/deloitte-acquires-gryphon-scientific-business-to-expand-security-science-and-public-health-capabilities.html (opens in a new window) 23 L. Weidinger, M. Rauh, N. Marchal, A. Manzini, L. A. Hendr... |
SOURCE_URL: https://www.anthropic.com/transparency |
FINAL_URL: https://www.anthropic.com/transparency |
HTTP_STATUS: 200 |
ARTEFACT_TYPE: public_docs |
ACCESSED: 2026-08-04 |
---------------------------------------------------------------------- |
End of preview. Expand in Data Studio
EU AI Act Article 50 Transparency Scoreboard
Version 1.0 (August 2026 Snapshot) · NM AI Research · CC BY 4.0 ORCID: 0009-0003-4213-7769 · DOI: 10.5281/zenodo.21819102
Dataset Summary
An empirical regulatory assurance dataset auditing 12 frontier AI consumer products and developer APIs against the European Union AI Act Article 50 Transparency Obligations, which took effect on 2 August 2026.
The dataset evaluates observable corporate artefacts (Terms of Service, System Cards, Transparency Notes, Technical Documentation, and Verification Tools) against an 8-item rubric derived from the European Commission Guidelines on Article 50 implementation (C(2026) 5054 final).
Evaluated AI Providers and Products
- Anthropic (Claude)
- OpenAI (ChatGPT)
- Microsoft (Copilot Consumer)
- Google (Gemini)
- Meta (Meta AI)
- Mistral AI (Le Chat)
- Cohere (Coral / Command)
- Aleph Alpha (Pharia)
- Black Forest Labs (Flux)
- Synthesia (Synthesia Video)
- Resemble AI (Resemble Audio)
- Getty Images (Generative AI by Getty)
Rubric Structure and Clauses
- Article 50(1): AI Interaction Transparency (identifying that the user is interacting with an AI system).
- Item 1.1: AI-nature disclosure in provider documentation or terms.
- Item 1.2: In-product disclosure at first point of interaction.
- Article 50(2): Synthetic Media Marking and Provenance.
- Item 2.1: Machine-readable marking or provenance signals (C2PA, SynthID, metadata).
- Item 2.2: Technical marking methodology documented publicly.
- Item 2.3: Public verification or detection route documented.
- Item 2.4: Comprehensive modality coverage across generated outputs.
- Article 50(5): Presentation, Timing, and Accessibility.
- Item 5.1: Clear and distinguishable presentation.
- Item 5.2: Accessibility standard commitment.
Dataset Files
article50_scores.csv: 98 audit observations with verbatim evidence URLs, HTTP status codes, SHA-256 body hashes, and adjudicator notes.article50_rubric_v2.md: The complete 8-item observable scoring protocol.reproduce.py: Python 3 standard library script to verify scores, check hash bundles, and render summary tables.
How to Reproduce
python3 reproduce.py
Conflict of Interest Disclosure
NM AI Research is an independent analyst practice. Research and tooling were assisted by AI coding models from Anthropic and Google DeepMind. In accordance with governance standards, all claims, scores, and evaluations are grounded in verifiable, public primary records without model-as-author attribution.
- Downloads last month
- 228