– Show Me The Prompt. https://
bit.ly/4cT6N7B
#AI #MachineLearning #DeepLearning #LLMs #DataScience
RESEARCH
-
Show Me The Prompt: Advanced LLM Prompt Engineering Guide
By
–
-
Attention Mechanisms for Neural Network Optimization at ICML2024
By
–
Drop by the #ICML2024 Google Research booth today during the 4pm coffee break when Thomas Fu will demonstrate how to use attention mechanisms to optimize neural network structures, such as feature pruning and embedding table optimization.
-
Llama 3.1 Model Sizes Frontier Trends Analysis
By
–
It’s not best fit or regression lines since different model sizes/capabilities are on the graph. Why would you best fit to the average of the new 8B, 70B and 405B llama 3.1 models? These are just rough trends lines of the frontier
-
Open-source Models Finally Beating Closed-source Models
By
–
One year ago, I created a video in which I said, hopefully open-source models are going to beat closed source models soon. I got this comment and a response that it was only possible in dreams. Today, I responded to the response 🙂
-
MMLU Benchmark Standardization Issues in AI Evaluation
By
–
Lack of standardization in the benchmarks. To be fair, MMLU is not that bad compared to many other evals
-
Non-autoregressive LLM methods: counterintuitive yet impressive
By
–
Claro, me refiero a que si ya los LLMs autorregresivo son impresionantes, estos métodos no autorregresivos además de ser impresionante se me antojan super poco intuitivos.
-
Diffusion Models Generate Code Non-Sequentially Like Impressive Breakthrough
By
–
Ver a una IA de forma autorregresiva generando código (secuencialmente, token tras token) ya era impresionante.
— Carlos Santana (@DotCSV) 24 juillet 2024
Pero ver métodos similares a los modelos de difusión generando trozos de código y llegando a una solución, además de impresionante ya me parece anti-intuitivo 😯 https://t.co/J3dFfUaP6pVer a una IA de forma autorregresiva generando código (secuencialmente, token tras token) ya era impresionante. Pero ver métodos similares a los modelos de difusión generando trozos de código y llegando a una solución, además de impresionante ya me parece anti-intuitivo
-
Google Research Recruitment Opportunities at ICML2024
By
–
Have questions about research, internship, and employment opportunities at Google? Stop by the #ICML2024 Google Research booth at 1pm CEST today to join recruiters Kelsie Pedone & Adrian Oprea, who will answer all your questions!
-
Multi-Task Model Training at Scale With Multiple Data Sources
By
–
How do you train, serve, & maintain a multi-task model with multiple clients controlling different data sources all at large scales? At 1:30pm CEST today, @smritichirps
, Andrew Gilchrist-Scott & Brad Stocks will be at the #ICML2024 Google booth to discuss our lessons learned.