To be fair the claimed novelty in the MSJ paper isn’t k-shot jailbreaking itself, it’s that tuning mitigations sufficient in the short-context regime are ineffective when k is large
MACHINE LEARNING
-
Ray Databricks Integration Now GA for AI Development
By
–
The integration of Ray and Databricks offers a uniquely powerful solution to your AI development needs. Now in GA, see how this collaboration creates a cohesive ecosystem where logical and data parallelism thrive together:
-
Jeevantika Lingalwar’s Journey Through Data Science and AI
By
–
Jeevantika Lingalwar’s Journey Through Data Science, AI, and Quantum Computing: We, at Anaytics Vidhya, believe in the power of storytelling. Celebrating Women’s Day, we… https://
analyticsvidhya.com/blog/2024/03/j
eevantika-lingalwars-journey-through-data-science-ai-and-quantum-computing/?utm_source=dlvr.it&utm_medium=twitter
… #DataAnalytics #DataScience #DataDriven #Blockchain #CIO #SaaS #DeepLearning #AI -
Promising AI Approach Gains Traction After Public Discussion
By
–
I enjoyed the HN conversation about it – looks promising to me https://
news.ycombinator.com/item?id=400676
77
… Love the clarity of the write-up. I expect we will see if the approach has legs within a month or two now that the information is out there -
Running AI Models on GPU: Learning Technical Implementation
By
–
The point of the messing around is to learn more about how to run models on a GPU
-
Gemini 1.5 Pro In-Context Learning Scaling Techniques
By
–
It's been quite interesting to study how Gemini 1.5 Pro scales its in-context learning (ICL) from few to many shots. I found our experiments that avoid using hand-labeled examples with Reinforced ICL, and Unsupervised ICL (i.e. shots are input examples only), particularly neat.
-
VASA-1: AI talking avatar from a photo
By
–
Ceci est un avatar "parlant" créé à partir uniquement d'une photo de la personne et qui agit comme un "masque".
— Defend Intelligence (Anis Ayari) (@DFintelligence) 18 avril 2024
En bref, vous mettez votre caméra, votre micro, vous choisissez le visage que vous voulez avoir et BOOM.
Il s'agit d'un modèle d'IA de Microsoft intitulé "VASA-1".… pic.twitter.com/RJSo8NWZOVThis is a "talking" avatar created solely from a photo of the person and acting as a "mask". In short, you turn on your camera, your microphone, choose the face you want, and BOOM. This is a Microsoft AI model called "VASA-1".
-
Current State of AI: Comprehensive Summary and Overview
By
–
A good summary on the current state of AI
-
Optimizing ML Algorithm Time Complexity for AI Projects
By
–
Decoding the time complexity of ML algorithms can optimize your AI projects! For more in-depth analysis, follow @ingliguori and boost your expertise with 'The Digital Edge' https://
bit.ly/3u4pILl #MachineLearning #DataScience #AI -
Training Strategy: Pre-2021 Data Priority Over AI-Generated Content
By
–
With all the AI-generated content flooding the web There might be something to first training on content from 2021-on And then continuing to train on pre-2021 content