On the latest "Grounded Truth" podcast episode with guest Shayan Mohanty, CEO of Watchful.io learn "A Surprisingly Effective Way to Estimate Token Importance in LLM Prompts" - YouTube: hubs.ly/Q024ymG_0 | Spotify: tinyurl.com/yc4uyuvu #LLMs #PromptEngineering #ChatGPT #AI

Watchful.io CEO Shayan Mohanty joins The TWIML AI Podcast in a discussion on the data labeling aspect of the machine learning process and ways that a data-centric approach could add value and reduce cost by multiple orders of magnitude. - hubs.ly/Q021x_FD0

Traditional #datalabeling is "dying" in a world of #LLMs. Watchful.io co-founder and Head of Success John Singleton speaks to how this shift is impacting the way companies annotate their data and train their #ml models going forward. #datascience #AI - hubs.ly/Q021y2s-0

Ep.9 Grounded Truth - "What is Prompt Ensembling?"-Shayan Mohanty of Watchful.io discusses the intricacies of embedding #LLMs using prompt chaining and prompt ensembling. YouTube: hubs.ly/Q021y1mR0 | Spotify: tinyurl.com/yc67fbu8 | Apple: tinyurl.com/ms77sx5

1/ Another week another research drop. This time, about measuring two different types of uncertainty from LLM responses! You can play with the demo here: uncertainty.demos.watchful.io We're psyched about this one because it gets us a few steps closer to directly comparing LLMs.

Traditional #datalabeling is "dying" in a world of #LLMs. Watchful.io co-founder and Head of Success John Singleton speaks to how this shift is impacting the way companies annotate their data and train their #ml models going forward. #datascience #AI - hubs.ly/Q021xT5f0

Ep.9 Grounded Truth - "What is Prompt Ensembling?"-Shayan Mohanty of Watchful.io discusses the intricacies of embedding #LLMs using prompt chaining and prompt ensembling. YouTube: hubs.ly/Q021xT5n0 | Spotify: tinyurl.com/yc67fbu8 | Apple: tinyurl.com/ms77sx5

Watchful.io CEO Shayan Mohanty joins The TWIML AI Podcast in a discussion on the data labeling aspect of the machine learning process and ways that a data-centric approach could add value and reduce cost by multiple orders of magnitude. - hubs.ly/Q021xQSk0

Learn how Wilson Sonsini is leveraging their world-class legal subject matter experts to effectively apply labels to classify their large body of data, with both a high degree of accuracy and at scale, using Watchful.io. #AI #ML - hubs.ly/Q021xy1n0

Ep.9 Grounded Truth - "What is Prompt Ensembling?"-Shayan Mohanty of Watchful.io discusses the intricacies of embedding #LLMs using prompt chaining and prompt ensembling. YouTube: hubs.ly/Q021y1m00 | Spotify: tinyurl.com/yc67fbu8 | Apple: tinyurl.com/ms77sx5

Traditional #datalabeling is "dying" in a world of #LLMs. Watchful.io co-founder and Head of Success John Singleton speaks to how this shift is impacting the way companies annotate their data and train their #ml models going forward. #datascience #AI - hubs.ly/Q021xLd-0

Learn how Wilson Sonsini is leveraging their world-class legal subject matter experts to effectively apply labels to classify their large body of data, with both a high degree of accuracy and at scale, using Watchful.io. #AI #ML - hubs.ly/Q021xPSg0

NEWS: Watchful Launches Gen AI Open-Source Tools, Pioneering Transparency for Large Language Models (LLMs) #AI #MachineLearning #LLMs Learn more at: hubs.ly/Q028l_G90

From diginomica: Watchful bridges generative AI transparency gap diginomica.com/watchful-bridg…

Watchful.io in the News: Unlocking Metrics for Black Box LLMs - AIwire hubs.ly/Q0298q960 We at Watchful are excited the #AI community is embracing our most recent research! Transparency is needed to truly unlock the power of Large Language Models #LLMs #ChatGPT

Ep.9 Grounded Truth - "What is Prompt Ensembling?"-Shayan Mohanty of Watchful.io discusses the intricacies of embedding #LLMs using prompt chaining and prompt ensembling. YouTube: hubs.ly/Q0298p-C0 | Spotify: tinyurl.com/yc67fbu8 | Apple: tinyurl.com/ms77sx5

🚀 New year, new research drop: outperforming GPT-4 in synthetic text generation! Our new approach combines geometry and latent space magic to produce data that is faithful to the real stuff, and orders of magnitude cheaper to produce. Check out the full paper here:

this was a neat little read. the idea is to do data augmentation with embeddings. they randomly sample around an embedding and then decode with vec2text. there is a trick to randomly sampling while not leaving the embedding manifold; they try to sample within an embedding "cone"

enjoyed reading this. useful insights into structures in latent space.

Watchful’s team and technology are now part of Thoughtworks ! We are excited to announce that our team of talented experts have been acquired by Thoughtworks, a global technology consultancy. Learn more: investors.thoughtworks.com/news-releases/…