In 2018, Torvalds decided to pull back from his angry young man stance. He took a break from the Linux kernel to work on his behavior toward other developers. After he got a handle on it, Torvalds returned to the kernel. He's been much more mild-tempered since then. As he
OPEN SOURCE
-
ToolBench: Evaluating LLM-Based API Agent Performance
By
–
Agent-based manipulation of APIs using #LLMs is a popular approach, but consistent and reliable evaluation metrics to assess this has been lacking. In this poster, we introduce a set of benchmarks called ToolBench and evaluate multiple open source #LLMs. @SambaNovaAI Researcher
-
LangChain Splits Into Core and Community Packages
By
–
Towards LangChain 0.1: LangChain-Core and LangChain-Community TL;DR: We've split the old langchain package into three separate packages to improve developer experience `langchain-core` contains simple, core abstractions that have emerged as a standard, as well as
-
Andrew McCallum’s OpenReview: Advancing AI Research Review
By
–
Hanging out with an old friend and one of the unsung heroes of AI research — @andrewmccallum has been running @openreviewnet for the past decade as a non-profit. The improvements that this brought to AI and to academic reviewing — including running reviewing at NeurIPS, ICML,
-
Mistral-Medium Generates Functional Game Code on First Attempt
By
–
Guys… Mistral-Medium is really good 😳
— Pietro Schirano (@skirano) 12 décembre 2023
It coded a functional game on the first try.
Sure, the game isn't perfect (I could have been more explicit in the prompt) but no other open-source model I've ever tested gave me a functional game on the first try.
Not even close. pic.twitter.com/ozmbk0vNFJGuys… Mistral-Medium is really good It coded a functional game on the first try. Sure, the game isn't perfect (I could have been more explicit in the prompt) but no other open-source model I've ever tested gave me a functional game on the first try. Not even close.
-
Google Introduces Croissant ML Dataset Format for Interoperability
By
–
Drop by the Google booth at 9:15am today where @laroyo
, Meg Risdal & Omar Benjelloun will introduce Croissant —in collaboration with @MLCommons & others—a format for ML datasets with the goal of making datasets interoperable across repos & platforms. -
Mistral leads Paris AI boom as open source divides big tech
By
–
Excellent NYT article about parisian AI unicorn Mistral, the effervescent AI ecosystem in Paris, the rapidly increasing strength of the open source AI community, the support of open source AI by Meta, and the opposition to it by Google and OpenAI.
-
GPT4 tested across multiple languages
By
–
And I maintain that saying this is simply false because since GPT2, they have also tested it in other languages. + There is an entire section in the GPT4 technical report that explains how they tested it, for example, on MMLU by translating it into other languages.
-
Gemini vs Mistral: Cost and Model Accessibility Tradeoffs
By
–
gemini announcement: cost google millions of dollars. blog post. script. video production. landing pages. can't use the good model. oh well mistral 8×7: magnet link only. no explanation of what it is. no explanation of how to run it. might power everything soon
