New paper from our team, led by Pat Verga
Are you:
* Doing evaluation with LLMs?
* Using a huge model?
* Worried about self-recognition?

Try an ensemble of smaller LLMs.
Use a PoLL: less biased, faster, 7x cheaper. Works great on QA & Arena-hard evals
arxiv.org/abs/2404.18796

⌘ R + is now my default translation tool. It does a much better job at handling the figures of speech and nuances :)

OpenAI got huge competition.

Command R+ by Cohere

Here are 6 incredible features of Command R+, you don't want to miss:

Thanks to everyone building with the cohere toolkit! Lovely to see it in the GitHub trending repos :)

It has been such a privilege to get to work with Sara. What she has built out of Cohere For AI is absolutely wild. Thanks for everything Sara :) excited for what this next year brings.

Very excited to see what people build with this at BUILD DAY :) info.cohere.ai/cohere-build-d…

Build and deploy your own chatbot with citations in <1 minute `docker run -e COHERE_API_KEY=‘$CO_API_KEY’ -p 8000:8000 -p 4000:4000 ghcr.io/cohere-ai/cohe…`. We’re also releasing a fast, easy to use python sandbox called terrarium. We’ll be contributing more features in the next

If the rise of LLMs caught you by surprise, here's your chance to get a preview of what's likely to be the next monumental jump in AI capabilities:

LLM-backed agents that use software tools

In this video, I'll walk you through the concepts and code of building an LLM-backed

Get our models wherever models are sold!

if you are building anything to production with LLMs, there is a good chance embeddings are used somewhere in your product. Try out this new model. I bet you see a noticeable performance improvement.

Nils Reimers and the cohere have been cooking :)

This dog has produced so much shareholder value.

Cohere's Rerank 3 transforms enterprise search with 4k context length multilingual support venturebeat.com/ai/coheres-rer…

Poised to exceed $45 billion by 2030, the legal tech sector is rapidly leveraging the value of GenAI with retrieval-augmented generation (RAG). From redefining efficiency across research, document management, horizon scanning, and more to innovations in multilingual applications

日本のディペロッパーの皆さん、ありがとうございます！ :)

Cohere Rerank is the Franks RedHot sauce of LLMs;

I put that shit on everything!

this single line of code will drastically improve the accuracy of pretty much any search.

And it will add next to no cost and next to no latency. I can think of no reason not to use it. Nils Reimers and the team really killed it with this one.

find the details here