Over the past few months at Mozilla.ai, we engaged with a number of organizations to learn how they are using language models in practice. We spoke with 35 organizations across sectors like finance, government, startups, and large enterprises. Our interviewees ranged from ML engineers to CTOs, capturing a diverse range of perspectives. Our interview summary notes for the 35 conversations amounted to 18,481 words (approximately 24,600 tokens), almost the length of a novella. To avoid confirmation bias and subjective interpretation, we decided to leverage language models for a more objective analysis of the data. By providing the models with the complete set of notes, we aimed to uncover patterns and trends without our pre-existing notions and biases. For this, we used Llama-3-8B-Instruct-Gradient-1048k by Meta and Gradient; Phi-3-medium-128k-instruct by Microsoft; and Qwen1.5-7B-Chat by Alibaba Cloud. To read the GenAI trends across 35 organizations, check out our latest learnings by Stefan French! #machinelearning #LLM #GenAI
Mozilla.ai’s Post
More Relevant Posts
-
🌟 Highlights from our Lisbon gathering! 🌟 What an incredible week it's been for our team at Mozilla.ai! We just concluded an unforgettable working week in the beautiful city of Lisbon, Portugal. Our time was filled with team bonding activities, including an exciting Scavenger Hunt, and strategic brainstorming sessions. We also had crucial discussions about our open-sourcing strategy that provided a key moment to align and collaborate on our product vision. A huge thank you to everyone who made this off-site possible and to our amazing colleagues for their energy, enthusiasm, and collaboration. We are excited for what's next! Aaron Gonzales, Julie V. Belião, Mario David Cariñana Abasolo, Morgan Alexander, Patrícia Martins (she/her), Sandra Antunes, Claudia Bertani, Santiago Martorana, Stefan French David Manzano-Macho, PhD Juliana Araújo Jane Silber Davide Eynard Kyle White Imtihan Ahmed Vicki Boykis
-
+1
To view or add a comment, sign in
-
Mozilla and EleutherAI brought together experts to discuss a critical question: How do we create openly licensed and open-access LLM training datasets and how do we tackle the challenges faced by their builders? Mozilla.ai's Senior Director of Product Innovation, Julie V. Belião, participated in the workshop along the many outstanding experts. Thank you to Kasia Odrozek, Stella Biderman, and Ayah Bdeir.
Sr. Director of Product Innovation @ Mozilla.ai | Executive Consultant, Product, Strategy, Operations
I recently had the privilege of attending the Open Dataset Convening, organized by EleutherAI and Mozilla in Amsterdam. This event was super insightful as we discussed the creation and sustainability of openly licensed and open-access LLM training datasets 👐 Together, we explored how these datasets could be built, how to make them sustainable, and ways to support their producers. Topics included the potential impact on the industry, ensuring inclusivity and equity, addressing copyright and ethical considerations, and managing data privacy rights. Our discussions also delved into developing legal and ethical frameworks, establishing collaborative structures, and ensuring access via cultural and linguistic diversity. This aligns with many of our goals at Mozilla.ai ✨ where we are committed to transparency, access, agency and inclusivity. For more insights into these discussions and their implications, check out the great article published by Kasia Odrozek and Stella Biderman 🙏 (special thanks 💙 to Ayah Bdeir for organizing and inviting us, and to Santiago Martorana for his support and coordination 🤗 ) https://lnkd.in/d_9jjajG
Dataset Convening: A community workshop on openly licensed LLM datasets
https://blog.mozilla.org/en/
To view or add a comment, sign in
-
Yesterday we kicked off our Team workweek in #Lisbon. We’re aiming to keep both knowledge and hydration on the high-end of the spectrum while we work on our #opensource components and platform. Lots of interesting news to follow soon…!
To view or add a comment, sign in
-
This year at MozFest House Amsterdam, we participated in Mozilla's Data Future Labs Showcase. David Manzano-Macho, PhD, our VP of Engineering, was a panelist alongside EM Lewis Jong, Ayah Bdeir, and Saskia Lensink. The panel featured spectacular work by Spawning, the Data Provenance Initiative, Imperial College London, and First Languages AI Reality. Masterfully presented by Mozilla's very own Miguel Morachimo, check out the video to see this year's winner! Shayne Longpre Cullen Miller Nataša Krčo Michael Running Wolf Mozilla Festival Mark Surman
Data Futures Lab Showcase - MozFest House Amsterdam
https://www.youtube.com/
To view or add a comment, sign in
-
It was a delight and an honor to participate in MozFest House Amsterdam. Thank you to all the organizers for bringing together such wonderful people and organizations together under one roof and in one shared space.
🎉 A heartfelt thank you to everyone who joined us in person and online at #MozFest! From insightful conversations and workshops to inspiring art installations and performances, your participation was the key to our success. You came together as artists, tech enthusiasts, journalists, academics, entrepreneurs, activists, and policymakers, in solidarity and togetherness, to take action for our digital future. We witnessed connections form, healthy debates ignite, and innovative ideas take flight. Your enthusiasm and engagement made MozFest House Amsterdam truly memorable. This cheers goes to you! ⤵️ 📽️ https://mzl.la/45HPOmo
MozFest House Amsterdam - Thank You!
https://vimeo.com/
To view or add a comment, sign in
-
Julie V. Belião, Senior Director of Product Innovation of Mozilla.ai at TAUS 2024 Conference in Rome. Thank you Rares Vasilescu for posting!
One of the major disconnects in terms of language and audience accessibility in current large language models systems, from Julie V. Belião at #TAUS2024 #TAUSRome2024 #TAUS #translation #localization #llm #ai
To view or add a comment, sign in
-
Mozilla.ai's Juliana Araújo and Stefan French present at The Linux Foundation AI_dev Conference in Paris. If you see them, say hello, and let's talk #opensource #AI !
To view or add a comment, sign in
-
Julie V. Belião, Mozilla.ai’s Senior Director of Product Innovation, will be speaking and moderating a panel tomorrow on Multilingual AI at the TAUS #AI Conference in Rome. Join us!
The final article for the topic for the Massively Multilingual AI Conference in Rome: "Towards a Truly Multilingual AI: Breaking the English Dominance" by Julie V. Belião from Mozilla.ai. Read more to discover how the dominance of English in tech and academia creates barriers for non-native speakers and perpetuates biases, a subject that will be discussed in-depth at the #TAUSRome2024. See you there! 🇮🇹
Towards a Truly Multilingual AI: Breaking the English Dominance
TAUS on LinkedIn
To view or add a comment, sign in
-
At this year's Mozilla Festival, David Manzano-Macho, PhD, our VP of Engineering, will participate as a panelist in Mozilla's Data Future Labs Showcase on Thursday, 13th of June. The Data Future Labs (DFL) Showcase highlights local builders around the world developing data tools and platforms that prioritize the needs and interests of their communities. This edition will be focused on shifting the narrative around what’s possible in data stewardship and Trustworthy AI in Europe and will include organizations like Imperial College London, Spawning, First Languages AI Reality, and the Data Provenance Initiative. #MozFest
Mozilla Festival 2024 Schedule
schedule.mozillafestival.org
To view or add a comment, sign in
-
Next week, Julie V. Belião, Senior Director of Product Innovation at Mozilla.ai, will be speaking at the TAUS Massively Multilingual AI Conference 2024 at the panel "No English Please: How to move towards a truly multilingual AI" on Wednesday, 19th of June. The panel will also feature Kareem Darwish from aiXplain, Christian Federmann from Microsoft, Gina Moape from Mozilla's Common Voice Initiative, and Lucie Gianola from France's Ministère de la Culture. English is the dominant language in Tech and AI research. Large Language Models are trained more than 90% on English language data. Inevitably, the LLMs look at the world through English filters. The focus of the session will be how to overcome this language bias and how to move towards a more democratic and inclusive world of AI. Thank you Jaap Van Der Meer and Anne-Maj van der Meer for the invitation. https://lnkd.in/gn6MTWc6
To view or add a comment, sign in
2,027 followers