Introducing: Cohere Command A+
We’ve created our most powerful LLM yet, optimized it to run on as little hardware as possible, and released it open-source for all.
It is critical for scientific integrity that we trust our measure of progress.
The @lmarena_ai has become the go-to evaluation for AI progress.
Our release today demonstrates the difficulty in maintaining fair evaluations on @lmarena_ai, despite best intentions.
Today @cohere is very excited to introduce Command A, our new model succeeding Command R+. Command A is an open-weights 111B parameter model with a 256k context window focused on delivering great performance across agentic, multilingual, and coding usecases. 🧵
Today, we’re launching early access for North!
Our all-in-one secure AI workspace platform combines LLMs, search, and agents into an intuitive interface that effortlessly integrates AI into your daily work to achieve peak productivity.
Introducing our latest AI search model: Rerank 3.5!
Rerank 3.5 delivers state-of-the-art performance with improved reasoning and multilingual capabilities to precisely search complex enterprise data like long documents, emails, tables, and code.
https://t.co/UGxclqGIPY
Extremely proud to share ✨ Aya Expanse ✨
We are a small lab, and this builds on years of dedicated research to connect the world with language. 🌎🌍🌏
A huge shoutout to the @CohereForAI@cohere team and our wider community for breakthroughs that ensure the world is seen.
Introducing ✨Aya Expanse ✨ – an open-weights state-of-art family of models to help close the language gap with AI.
Aya Expanse is both global and local. Driven by a multi-year commitment to multilingual research.
https://t.co/0WsC2i9C8a
Today @CohereForAI and @cohere are releasing two new multilingual models spanning 23 popular languages. Across a range of use-cases, these models represent a new state-of-the-art.
We're also releasing the weights (8B & 32B) on @huggingface
As enterprises scale their use of large language models (LLMs), doing so securely is crucial.
We frequently work with customers to address these security challenges and ensure safe AI adoption. I highly recommend a read!
Scaling LLMs? Security is critical. Private LLM environments let you tap into AI’s power while protecting data.
Our guide covers safe deployment strategies, focusing on privacy, access control, and secure infrastructure. Devs, check it out!
Download now: https://t.co/FTIehaBYSF
The average cost of data breaches rose 10% in 2023 (IBM). As AI transitions to enterprise-grade, cybersecurity is vital.
Steve Wilson from Exabeam highlights some key AI security challenges: prompt injection, data poisoning, safe data practices, and regulatory prep. Industry collaboration is crucial. Read more: https://t.co/a94mUJxgxB
Introducing Expedition Aya 🌍
A 6-week open build challenge to accelerate multilingual progress.
Join us to expand the world that AI sees.
https://t.co/QmGpdTN5v4
Financial RAG Evaluation🕵️
Haiku got lots of excitement yesterday.
I ran it through my financial RAG eval pipeline today.
Haiku was fast, but struggled on correctness versus similar models.
Cmd-r remained financial RAG champ.
Main takeaways:
• haiku faster than gpt-3.5
• haiku less correct than gpt-3.5
• cmd-r fastest and most correct
2.83 seconds for retrieval + reranking + answering is excellent from @cohere
Experiment setup:
• 100 questions from 10-K
• cosine similarity retrieval
• post-retrieval reranking
• improved prompts
• ragas for correctness evals
Next step is to test more models in this tier like gemini 1.0 pro and groq's mistral 8x7b.
Also, need to optimize chunking.
However, have yet to find compelling approaches for chunking 10-Ks.
Suggestions welcome.
⌘-R
Introducing Command-R, a model focused on scalability, RAG, and Tool Use. We've also released the weights for research use, we hope they're useful to the community!
https://t.co/0Q4Uj6zFjd
Today, we’re launching Aya, a new open-source, massively multilingual LLM & dataset to help support under-represented languages. Aya outperforms existing open-source models and covers 101 different languages – more than double covered by previous models.
https://t.co/0WsC2i9C8a
There’s just 1 week until @emnlpmeeting! We’re looking forward to attending EMLNLP at the Resorts World Convention Centre in Singapore ✈️🇸🇬
Here’s where you can find the @Cohere & C4AI team throughout the conference, be sure to stop by our booth to say hello & grab some swag!
This threat report list is 🔥. As an intel analyst, I care most about the threats that are actually exploiting this vulnerability so we can help defenders prioritize. Please contribute if you can! #log4j#Log4Shell
The @DianaInitiative is officially one week away! I will be delivering my keynote, Building Security Champions, on July 16th at 11:00 am PDT!! Have you bought your tickets yet? 🤔
To learn more, click the link ⬇️
https://t.co/VtcBmPcwYi
Our recent blog on the #Kaseya#ransomware attack takes a closer look at the #malware and how other hackers are looking to ride the coattails of #REvil: https://t.co/ts03nclHVE
It’s #CyberMentoringMonday!!!! Are you looking for a professional mentor or to learn more about InfoSec? Are you experienced and willing to ‘give back’? Use this thread and hashtag to connect! 🎉