LLMS

→ Voir le post original sur X — @stephanemallard

OpenAI CEO Threatens to Exit Europe Over Restrictive AI Laws

By

@stephanemallard

–

25 May 2023 21h20

Le business des VPN va EXPLOSER "ChatGPT : le patron d'OpenAI menace de quitter l'Europe si la législation communautaire est trop restrictive" https://
lefigaro.fr/secteur/high-t
ech/face-a-la-mefiance-anti-ia-le-createur-de-chatgpt-en-tournee-mondiale-20230525
…

→ View original post on X — @cerebras

Cerebras Model Lab: Compare LLMs, Estimate Training Costs

By

@cerebras

–

25 May 2023 20h12

Introducing Model Lab: a new tool we built to make sense of training LLMs! The app lets you easily: – compare different models (LLaMA, Pythia etc)
– estimate training FLOPS, memory, inference cost
– visualize loss curves
Give it a try here https://
cerebras.net/model-lab/

→ View original post on X — @hwchase17

Enterprise LLM Adoption: Vertex AI and Azure AI Features

By

@hwchase17

–

25 May 2023 19h31

Similar to how AzureAI hosts OpenAI models but adds in enterprise features, this is what Vertex does with PaLM These enterprise-grade features are crucial for enabling enterprises to adopt LLMs. AzureAI has sneaky been a very popular integration for this exact reason

Are Language Models Training Us Instead of We Training Them?

By

–

25 May 2023 19h10

Are we training the language models, or are the language models training us?

Language Model Performance: Emotional Intelligence vs Reasoning Strategies

By

–

25 May 2023 19h10

This battle is quite revealing. Karina's prompting strategy was getting the language model to say it understands (emotional intelligence). My strategy was to get the model to reason more (IQ). Our strategies each worked on the respective language models that we are used to using.

Failed Claude Prompt Works Successfully on GPT-4

By

–

25 May 2023 19h10

Then i copy-paste in my prompt that failed on Claude. Much to my relief, it actually works on GPT-4.

Confirming GPT-4 Limitation with Simple Prompting Approach

By

–

25 May 2023 19h10

First, I confirm that it doesn't work on GPT-4 with simple prompting, to prove the prompt matters, and it's not just because GPT-4 is smarter than Claude.