Corpus / Changes
What changed, and when
Every version of every watched document is kept, so a document that stops saying what it used to say can be shown saying both. This is the output the corpus exists for, and the one thing here that cannot be reconstructed later by anybody who did not keep the earlier version.
| Changed | Document | What changed | |
|---|---|---|---|
| 2026-11-04 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 6 lines added, 6 lines removed | Compare → |
| 2026-11-04 | Introducing Shieldstral. | Mistral https://mistral.ai/news/shieldstral/ | 2 lines added, 3 lines removed | Compare → |
| 2026-11-03 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 7 lines added, 7 lines removed | Compare → |
| 2026-11-02 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-11-02 | Responsibility & Safety — Google DeepMind https://deepmind.google/responsibility-and-safety/ | 4 lines added, 4 lines removed | Compare → |
| 2026-11-01 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-31 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-31 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 4 lines added, 4 lines removed | Compare → |
| 2026-10-31 | https://huggingface.co/prism-ml/Ternary-Bonsai-2-27B-gguf | 24 lines added, 7 lines removed | Compare → |
| 2026-10-30 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 8 lines added, 8 lines removed | Compare → |
| 2026-10-29 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 6 lines added, 6 lines removed | Compare → |
| 2026-10-28 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-27 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-27 | AISI Research & Publications | The AI Security Institute https://www.aisi.gov.uk/research | 20 lines added | Compare → |
| 2026-10-26 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 6 lines added, 6 lines removed | Compare → |
| 2026-10-26 | Developer Use Guide for Llama https://developer.meta.com/ai/developer-use-guide/ | 1 line removed | Compare → |
| 2026-10-26 | https://huggingface.co/CMSManhattan/JiRackUltra_1b | 3 lines added | Compare → |
| 2026-10-25 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-24 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 5 lines added, 5 lines removed | Compare → |
| 2026-10-23 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-23 | Updates - METR https://metr.org/blog/ | 8 lines added | Compare → |
| 2026-10-22 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-22 | Generative AI Prohibited Use Policy https://policies.google.com/terms/generative-ai/use-policy | 1 line added | Compare → |
| 2026-10-22 | Usage Policy \ Anthropic https://www.anthropic.com/legal/aup | 5 lines added, 4 lines removed | Compare → |
| 2026-10-21 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-21 | Updates - METR https://metr.org/blog/ | 8 lines added | Compare → |
| 2026-10-21 | Anthropic’s Responsible Scaling Policy \ Anthropic https://www.anthropic.com/responsible-scaling-policy | 3 lines removed | Compare → |
| 2026-10-21 | Anthropic’s Transparency Hub \ Anthropic https://www.anthropic.com/transparency | 852 lines added, 212 lines removed | Compare → |
| 2026-10-20 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 4 lines added, 4 lines removed | Compare → |
| 2026-10-20 | Usage Policy \ Anthropic https://www.anthropic.com/legal/aup | 2 lines removed | Compare → |
| 2026-10-20 | Anthropic’s Transparency Hub \ Anthropic https://www.anthropic.com/transparency | 29 lines added, 29 lines removed | Compare → |
| 2026-10-19 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-18 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-17 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-16 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 5 lines added, 5 lines removed | Compare → |
| 2026-10-16 | Gemma 3 model card | Google AI for Developers https://ai.google.dev/gemma/docs/core/model_card_3 | 1 line added, 1 line removed | Compare → |
| 2026-10-16 | Gemma Prohibited Use Policy | Google AI for Developers https://ai.google.dev/gemma/prohibited_use_policy | 1 line added, 1 line removed | Compare → |
| 2026-10-15 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 7 lines added, 7 lines removed | Compare → |
| 2026-10-15 | Science – Apollo Research https://www.apolloresearch.ai/science | 7 lines added, 7 lines removed | Compare → |
| 2026-10-14 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 4 lines added, 4 lines removed | Compare → |
| 2026-10-14 | Responsibility & Safety — Google DeepMind https://deepmind.google/responsibility-and-safety/ | 8 lines added, 8 lines removed | Compare → |
| 2026-10-14 | Developer Use Guide for Llama https://developer.meta.com/ai/developer-use-guide/ | 1 line added, 1 line removed | Compare → |
| 2026-10-14 | https://huggingface.co/CMSManhattan/JiRackUltra_1b | 15 lines added, 6 lines removed | Compare → |
| 2026-10-14 | Model cards — Google DeepMind https://deepmind.google/models/model-cards | 1 line added | Compare → |
| 2026-10-13 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-12 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-11 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-10 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-09 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 6 lines added, 6 lines removed | Compare → |
| 2026-10-09 | Usage Policy \ Anthropic https://www.anthropic.com/legal/aup | 142 lines added, 94 lines removed | Compare → |
| 2026-10-08 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-08 | AISI Research & Publications | The AI Security Institute https://www.aisi.gov.uk/research | 10 lines added | Compare → |
| 2026-10-08 | Developer Use Guide for Llama https://developer.meta.com/ai/developer-use-guide/ | 12 lines added, 47 lines removed | Compare → |
| 2026-10-07 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-07 | Developer Use Guide for Llama https://developer.meta.com/ai/developer-use-guide/ | 8 lines added, 15 lines removed | Compare → |
| 2026-10-07 | Updates - METR https://metr.org/blog/ | 8 lines added | Compare → |
| 2026-10-06 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-06 | Introducing Shieldstral. | Mistral https://mistral.ai/news/shieldstral/ | 1 line added, 1 line removed | Compare → |
| 2026-10-05 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 4 lines added, 4 lines removed | Compare → |
| 2026-10-05 | https://huggingface.co/CMSManhattan/JiRackUltra_1b | 3 lines added | Compare → |
| 2026-10-04 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-03 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-03 | Anthropic’s Transparency Hub \ Anthropic https://www.anthropic.com/transparency | 62 lines added, 2 lines removed | Compare → |
| 2026-10-02 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 6 lines added, 6 lines removed | Compare → |
| 2026-10-02 | https://huggingface.co/CMSManhattan/JiRackUltra_1b | 1 line added, 1 line removed | Compare → |
| 2026-10-01 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-10-01 | AISI Research & Publications | The AI Security Institute https://www.aisi.gov.uk/research | 10 lines added | Compare → |
| 2026-10-01 | Updates - METR https://metr.org/blog/ | 8 lines added | Compare → |
| 2026-09-30 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 5 lines added, 5 lines removed | Compare → |
| 2026-09-30 | Introducing Shieldstral. | Mistral https://mistral.ai/news/shieldstral/ | 2 lines added, 2 lines removed | Compare → |
| 2026-09-30 | https://huggingface.co/CMSManhattan/JiRackUltra_1b | 4 lines added | Compare → |
| 2026-09-29 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 5 lines added, 5 lines removed | Compare → |
| 2026-09-28 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-09-27 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-09-26 | Data on AI models | Epoch AI https://epoch.ai/data/ai-models | 14 lines added, 14 lines removed | Compare → |
| 2026-09-25 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 3 lines added, 3 lines removed | Compare → |
| 2026-09-25 | https://huggingface.co/prism-ml/Ternary-Bonsai-2-27B-gguf | 1 line added, 1 line removed | Compare → |
| 2026-09-25 | Developer Use Guide for Llama https://developer.meta.com/ai/developer-use-guide/ | 6 lines added | Compare → |
| 2026-09-24 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 7 lines added, 7 lines removed | Compare → |
| 2026-09-24 | Responsibility & Safety — Google DeepMind https://deepmind.google/responsibility-and-safety/ | 4 lines added, 4 lines removed | Compare → |
| 2026-09-24 | https://huggingface.co/prism-ml/Ternary-Bonsai-2-27B-gguf | 23 lines added, 18 lines removed | Compare → |
| 2026-09-24 | Developer Use Guide for Llama https://developer.meta.com/ai/developer-use-guide/ | 2 lines removed | Compare → |
| 2026-09-24 | GPT-5.6 System Card - OpenAI Deployment Safety Hub https://deploymentsafety.openai.com/gpt-5-6/evaluations-with-challenging-prompts | 2630 lines replaced 2630 — too large to align line by line | Compare → |
| 2026-09-23 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-09-23 | Developer Use Guide for Llama https://developer.meta.com/ai/developer-use-guide/ | 5 lines removed | Compare → |
| 2026-09-23 | Updates - METR https://metr.org/blog/ | 8 lines added | Compare → |
| 2026-09-23 | Science – Apollo Research https://www.apolloresearch.ai/science | 2 lines added | Compare → |
| 2026-09-23 | GPT-5.6 System Card - OpenAI Deployment Safety Hub https://deploymentsafety.openai.com/gpt-5-6/evaluations-with-challenging-prompts | 2630 lines replaced 2575 — too large to align line by line | Compare → |
| 2026-09-22 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 5 lines added, 5 lines removed | Compare → |
| 2026-09-21 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-09-21 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 3 lines added, 3 lines removed | Compare → |
| 2026-09-21 | L_2023165EN.01000101.xml https://eur-lex.europa.eu/eli/reg/2023/1230/oj/eng | 5314 lines replaced 5596 — too large to align line by line | Compare → |
| 2026-09-20 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 4 lines added, 4 lines removed | Compare → |
| 2026-09-20 | Science – Apollo Research https://www.apolloresearch.ai/science | 1 line added, 1 line removed | Compare → |
| 2026-09-19 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-09-18 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 5 lines added, 5 lines removed | Compare → |
| 2026-09-18 | Anthropic’s Transparency Hub \ Anthropic https://www.anthropic.com/transparency | 52 lines added, 1 line removed | Compare → |
| 2026-09-17 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 6 lines added, 6 lines removed | Compare → |
| 2026-09-16 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 4 lines added, 4 lines removed | Compare → |
| 2026-09-15 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-09-14 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-09-13 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-09-12 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-09-11 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-09-11 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 3 lines added, 3 lines removed | Compare → |
| 2026-09-11 | Science – Apollo Research https://www.apolloresearch.ai/science | 35 lines added, 21 lines removed | Compare → |
| 2026-09-10 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 2 lines added, 2 lines removed | Compare → |
| 2026-09-10 | AILuminate - MLCommons https://mlcommons.org/benchmarks/ailuminate/ | 1 line added, 7 lines removed | Compare → |
| 2026-09-09 | Data on AI Models | Epoch AI https://epoch.ai/data/ai-models | 7 lines added, 7 lines removed | Compare → |
What this page will and will not tell you
- It reports size, never meaning. A change is described as lines added and removed. Calling an edit “a weakened safety commitment” is a judgement, and the diff has no business making one — read the text and decide.
- HTML pages are compared as their readable text. Navigation and build tokens move on every request, so comparing raw bytes would report a change every day. The fingerprint of the original bytes is kept alongside.
- A very large rewrite is reported coarsely. Beyond a few thousand changed lines the two versions are shown whole rather than aligned, because a line-by-line alignment of two unrelated documents is noise presented as precision.