nah that’s not the same thing at all; logit lens is pretty poorly conditioned esp at early layers and certainly doesn’t capture all of the hidden state
LLMS
-
LangSmith General Availability Launch on Product Hunt
By
–
Give us some love on Product Hunt for our GA launch of LangSmith today! Link here: https://
producthunt.com/posts/langsmit
h-general-availability?bc=1
… -
Latent Text Transformer: Making Model Thoughts Readable
By
–
random research idea: Latent Text Tansformer (LTT) in a nutshell: replace sequence of *vectors* as hidden states of the transformer with sequences of *tokens*, so we can read the model's "thoughts" directly then train a transformer that uses longer sequences of *discrete
-
OpenAI reportedly using GPT-5 as research assistant
By
–
Totally agree with that. But I think it's not a far-fetched scenario to imagine OpenAI at this point has been using GPT-5 in some form, and maybe as a "research assistant" in this case.
-
LangSmith Platform Overview for LLM Application Development
By
–
Check out our in-depth platform overview of LangSmith if you've missed our webinar today: https://
youtube.com/watch?v=3wAON0
Lqviw
… LangSmith is an all-in-one developer platform for every step of the LLM-powered application lifecycle – whether you’re building with LangChain or not. LangSmith -
Groq Recipes with Sweet Baby Rays Demo Invitation
By
–
Hey @finkd
! Have you had a chance to try out the recipes by Groq® that use #SweetBabyRays yet? Either way, you have to admit the demo is pretty . You should try it. https://
hubs.ly/Q02knhFh0 #Groq @Meta #Llama2 #LLM #GenAI -
Gemini 1.5 Announcement Overshadowed by OpenAI’s Dominance
By
–
It was the announcement of Gemini 1.5 released today with More and multimodality. But tbh: no one cares, openAI blasted it
-
GPT-5 and AI video expected to shock world in 2026
By
–
Yep. Gpt-5 and AI video are going to shock the world this year I believe
-
Major AI Announcements: Sora, Gemini, Magic.dev Raises
By
–
Bad day to take a long lunch. Crazy day for AI announcements. – OpenAI Sora
– Google Gemini 1.5
– http://
Magic.dev $100m raise
– Meta V-JEPA
– Lambda Labs $320m series C -
OpenAI announces new Text to video model: Sora
By
–
OpenAI’s new Text to video model “Sora” just announced pic.twitter.com/fvhcwxc6qb
— AI Breakfast (@AiBreakfast) 15 février 2024OpenAI’s new Text to video model “Sora” just announced