2024-02-26 23:45:22
Well that is very interesting! Introducing Mistral-Large on Azure in partnership with Mistral AI | Microsoft Azure Blog
https://azure.microsoft.com/en-us/blog/microsof…
Well that is very interesting! Introducing Mistral-Large on Azure in partnership with Mistral AI | Microsoft Azure Blog
https://azure.microsoft.com/en-us/blog/microsof…
Multi-purpose InSTRument for Astronomy at Low-resolution: MISTRAL@OHP
J. Schmitt, C. Adami, M. Dennefeld, F. Agneray, S. Basa, J. C. Brunel, V. Buat, D. Burgarella, C. Carvalho, G. Castagnoli, N. Grosso, F. Huppert, C. Moreau, F. Moreau, L. Moreau, E. Muslimov, S. Pascal, S. Perruchot, D. Russeil, J. L. Beuzit, F. Dolon, M. Ferrari, B. Hamelin, A. LevanSuu, K. Aravind, D. Gotz, E. Jehin, E. LeFloc'h, J. Palmerio, A. Saccardi, B. Schneider, F. Sch\"ussler, D. Turpin, S. D. Verg…
KI-Update kompakt: Mistral, Nvidia RTX, KI-Konzepthandy, Umweltkosten von KI
Das "KI-Update" liefert werktäglich eine Zusammenfassung der wichtigsten KI-Entwicklungen.
https://www.
Patronus AI releases CopyrightCatcher and says GPT-4 produced copyrighted content on 44% of prompts, Mistral's Mixtral on 22%, Llama 2 10%, and Claude 2.1 8% (Hayden Field/CNBC)
https://www.cnbc.com/2024/03/06/gpt-4-rese
Simple Techniques for Enhancing Sentence Embeddings in Generative Language Models
Bowen Zhang, Kehua Chang, Chunping Li
https://arxiv.org/abs/2404.03921 ht…
Microsoft says its investment in Mistral AI amounts to €15M; EU regulators plan to analyze the investment, which could eventually lead to a formal investigation (Samuel Stolton/Bloomberg)
https://www.bloomberg.com/news/articles/20
https://techcrunch.com/2024/02/26/mistral-ai-releases-new-model-to-rival-gpt-4-and-its-own-chat-assistant/?eicker.news
In December I tried the small model. Really good, also in ressource consumption.…
SaulLM-7B: A pioneering Large Language Model for Law
Pierre Colombo, Telmo Pessoa Pires, Malik Boudiaf, Dominic Culver, Rui Melo, Caio Corro, Andre F. T. Martins, Fabrizio Esposito, Vera L\'ucia Raposo, Sofia Morgado, Michael Desa
https://arxiv.org/abs/2403.03883
KI-Update kompakt: Gemini 1.5, Mistral Next, Altmans Chip-Pläne, Copilot
Das "KI-Update" liefert werktäglich eine Zusammenfassung der wichtigsten KI-Entwicklungen.
https://www.
Microsoft partners with Mistral as regulators probe its OpenAI alliance; Microsoft takes a minor stake in Mistral and help bring Mistral's AI models to market (Madhumita Murgia/Financial Times)
https://t.co/9Jk7fp65JN
This https://arxiv.org/abs/2403.03883 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCL_…
#AI at it‘s best …
@… yeah I know, thanks. I already switched to gpt4, I only ment: I still have to test the Mistral model :)
Elon Musk might be suing OpenAI because he is a troll and has a grudge but his lawsuit puts a spotlight on the empty humanitarian posturing by AI businesses (Parmy Olson/Bloomberg)
https://www.bloomberg.com/opinion/articles
Workshopfrage letzte Woche: "Wie kann ich selber ein offenes LLM z.B auf meinem Laptop austesten, wenn ich keine GPU habe?"
⏩ Ollama.ai kann Modelle auch nur mit einer CPU ausführen. Das genügt zum Testen (Output wird langsam generiert), man braucht nur Festplattenplatz.
Wer ein LLM auf Deutsch testen möchte, kann dann einfach > ollama run mixtral ausführen. Über die Modelfile-Konfigurationsdatei lässt sich via System Prompt einstellen, wie das Modell antworten soll…
#AI at it‘s best …
Source: Mistral is in talks with investors about raising several hundred million dollars at a $5B valuation, after raising $415M at a $2B valuation in Dec. 2023 (Kate Clark/The Information)
https://www.theinformation.com/articles/mi
Introducing Super RAGs in Mistral 8x7B-v1
Ayush Thakur, Raghav Gupta
https://arxiv.org/abs/2404.08940 https://arxiv.org/pdf/2404.0894…
Nick Clegg, who could have insisted on PR instead of FPTP, but instead enabled Cameron to be an awful PM who caused Brexit, is now talking nonsence about AI
https://apnews.com/article/meta-ai-assistant-llama3-large-lan…
New releases for the major AI models.
https://www.theguardian.com/technology/2024/apr/10/ai-race-heats-up-as-openai-google-and-mistral-release-new-models
A profile of Mistral AI CEO Arthur Mensch, who says, as an atheist, he is uncomfortable with Silicon Valley's "AGI rhetoric" and "religious" fascination with AI (New York Times)
…
Sentinels of the Stream: Unleashing Large Language Models for Dynamic Packet Classification in Software Defined Networks -- Position Paper
Shariq Murtuza
https://arxiv.org/abs/2402.07950
We review the high end Wolf Coffee machine. We examine why Tyler Perry is right to embrace AI entertainment. We compare API friendliness for Anthropic, Gemini, Mistral and Perplexity!
Here's the question we asked all our API Bots in our test:
"Where do you start to write a story? Explain your process. As well, tell us why the name 'Elara' is so often used by LLMs in creative storytelling as a name."
A profile of Mistral AI CEO Arthur Mensch, who says, as an atheist, he is uncomfortable with Silicon Valley's "AGI rhetoric" and "religious" fascination with AI (New York Times)
…
Meta claims both Llama 3 models beat similarly sized models like Gemini, Mistral, and Claude 3 in certain benchmarks; humans marked Llama 3 higher than GPT-3.5 (Emilia David/The Verge)
https://www.theverge.com/2024/4/18/24134103/llama-…
Verif.ai: Towards an Open-Source Scientific Generative Question-Answering System with Referenced and Verifiable Answers
Milo\v{s} Ko\v{s}prdi\'c, Adela Ljaji\'c, Bojana Ba\v{s}aragin, Darija Medvecki, Nikola Milo\v{s}evi\'c
https://arxiv.org/abs/2402.18589
@… yes, Claude is amazing as well, but phind is able to search the web (like Perplexity.ai) - that's a real benefit. When it comes to coding questings, phind - my opinion - shows better results than gpt4, Mistral and Perplexity.ai
Sources: Mistral plans to raise €500M at an over €5B valuation from multiple VCs and wealth funds, after raising ~€400M at a €2B valuation in December 2023 (Financial Times)
https://t.co/aldaMDD3Jz
This https://arxiv.org/abs/2402.12147 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCL_…
Do Large Language Models Learn Human-Like Strategic Preferences?
Jesse Roberts, Kyle Moore, Doug Fisher
https://arxiv.org/abs/2404.08710 https://
Mistral AI launches Mixtral 8x22B, its latest sparse mixture-of-experts model, after releasing Mixtral 8x7B in December 2023 (Shubham Sharma/VentureBeat)
https://venturebeat.com/ai/mistral-ai-drops-new-mixture-of-experts-model-with-a-t…
This https://arxiv.org/abs/2402.10693 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCL_…
https://www.youtube.com/watch?v=3iZ7jwbAwF8
Embedding Layer of mistral-7b and gemma-7b
Chris Hay plays with the models. Good practical insights with hands on using python.
Interessant, bij #Perplexity in een labs omgeving de wat kleinere versies van de bekende modellen online beschikbaar, Claude, Gemma, Mistral etc.
https://labs.perplexity.ai/
This https://arxiv.org/abs/2402.13116 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCL_…
Mistral.rs: A Lightning-Fast #LLM #Inference Platform with Device Support, Quantization, and #OpenAI #API
The GPT-4 barrier has finally been broken, with Gemini 1.5, Mistral Large, Claude 3 Opus, and Inflection-2.5 benchmarking near or even above OpenAI's model (Simon Willison's Weblog)
https://simonwillison.net/2024/Mar/8/gpt-4-barrier/
This https://arxiv.org/abs/2402.13116 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCL_…