ππ²πΊππππΆπ³ππΆπ»π΄ #ArtificialIntelligence – 20 #AI Concepts explainedΒ #MachineLearning #ML #innovation #ChatGPT #chatbots #chatgpt4 #chatgpt3 #GenerativeAI #AIEthics Source Dirk Zee @mikeflache @pascal_bornet @Ronald_vanLoon @MargaretSiegien @BetaMoroney
LLMS
-
Model not found on Fireworks AI platform list
By
–
I can't find it on the list on https://
fireworks.ai/models -
Together.ai offers LLM API at $1.20 per million tokens
By
–
Looks like https://
together.ai have it for $1.20/million tokens -
Which LLM API Providers Offer the New Mixtral Model?
By
–
Which of the LLM API providers have the new Mixtral model?
-
Poe Updates GPT-4 Turbo to Improved Version
By
–
Weβve updated GPT-4 and GPT-4 128k in Poe to the new, greatly improved version of GPT-4 Turbo!
-
Mistral Releases Mixtral 8Γ22B Frontier LLM Model
By
–
French AI startup Mistral released Mixtral 8Γ22B, a powerful new frontier LLM, dropped quietly via a 281GB file on X available for download. The LLM features a 65,000-token context window, 176B parameters and is expected to surpass the previous Mixtral.
-
Prompt Engineering Tricks: Context Shapes LLM Math Performance
By
–
Pro-tip: remember that if you want Llama 2 to do well on a 50 question math test, pretend you are in Star Trek. To solve 100 questions, pretend you are in a political thriller. Results like this are why I stopped worrying about optimal prompting too much. https://
oneusefulthing.org/p/captains-log
-the-irreducible-weirdness
β¦ -
Grok headline generation raises AI transparency and bias concerns
By
–
The Grok generated headlines in Twitter's Explore feature look reasonable but the source tweet for the neutral headline is often an extreme or controversial account If there is really a commitment to AI transparency, it would be good to see how the AI is picking & shaping these
-
LLM as Judge: Are All Results Mediocre 7 out of 10?
By
–
now the bone chilling question for the lazy people who use llm as judge for everything: have you run this analysis on your judging results and is everything ~7 out of 10
-
Instructing AI Models: Treat Code as a Desired Response Format
By
–
The trick is never to say βDo X, using code.β Instead, say, βIn a code block, do X.β Donβt treat code as a tool it can decide when to use β treat it as your desired response format, so all thinking happens in comments. (UI bonus: rationales now fold/hide with code.)