Oh I didn't know that – very cool. I wonder if they're using our repetition tokens too.
LLMS
-
GCP pourrait dominer avec des modèles de 100M+ tokens
By
–
Nothing to test here but a big win for GCP if it is rly good At some point nobody will care about context windows though b/c every model will be capable to run 100M+
-
Chat RAG: Interactive Coding Assistant with Gradio and Ollama
By
–
Chat RAG Interactive Coding Assistant github: https://
github.com/JakeFurtaw/Cha
t-RAG
… Uses a @Gradio App to chat with any model available to @ollama and any files you load into the data directory -
Building Community Together with Llama Developers
By
–
We're building this community together — glad to have you along for the journey and grateful for everything you do for developers building with Llama!
-
Meta thanks NVIDIA for supporting Llama ecosystem
By
–
Huge thank you to the @NVIDIA team for everything you do to support the Llama ecosystem!
-
GPT-4o Token Prices Drop 89% in 17 Months
By
–
After a recent price reduction by OpenAI, GPT-4o tokens now cost $4 per million tokens (using a blended rate that assumes 80% input and 20% output tokens). GPT-4 cost $36 per million tokens at its initial release in March 2023. This price reduction over 17 months corresponds to
-
Qwen2-VL Released: New Vision Language Model Available
By
–
Qwen2-VL is out demo: https://
huggingface.co/spaces/Qwen/Qw
en2-VL
…
collection: https://
huggingface.co/collections/Qw
en/qwen2-vl-66cee7455501d7126940800d
… -
Nine Amazing AI Tools: Google GPT Alternative and Coding Innovation
By
–
Dive into 9 Amazing AI Tools This Week! From Google's custom GPT alternative to a viral AI coding tool, there's a lot to uncover. Here’s a quick rundown! #AI #Innovation #Tech #Futurepedia
-
Writing Custom LLM Benchmarks with Nicholas Carlini
By
–
🆕 Why you should write your own LLM benchmarks
— Latent.Space (@latentspacepod) 29 août 2024
w/ Nicholas Carlini of @GoogleDeepMind
Covering his greatest hits:
– How I Use AI
– My benchmark for large language models
– Extracting Training Data from Large Language Models (RIP @openai logprobs)
Full episode below! pic.twitter.com/TtVkNyIa9cWhy you should write your own LLM benchmarks w/ Nicholas Carlini of @GoogleDeepMind Covering his greatest hits:
– How I Use AI
– My benchmark for large language models
– Extracting Training Data from Large Language Models (RIP @openai logprobs) Full episode below! -
Llama Reaches 350M Downloads on Hugging Face Hub
By
–
"Llama is approaching 350M downloads on Hugging Face More than 10x compared to this time last year"