It appears to me from today and I saw more reports recently. It could be gradual though. I didn't notice it switching to o1 yet
LLMS
-
User reports updated knowledge cutoff in AI search features
By
–
I have a custom prompt to print a knowledge cutoff after every post. Since yesterday, I have seen an updated date of 09.2024, but only when I use the forced search feature Have you noticed it too? https://
x.com/kondrich2/stat
/kondrich2/status/1835821248422768760
… -
ChatGPT Rolls Out Auto-Model Selection Feature to All Users
By
–
BREAKING: ChatGPT seems to be rolling out "auto" mode to all users and all platforms. Auto mode used to be in limited testing for quite a while. This dynamic mode selects the best suitable model depending on the prompt.
-
ColPali: Efficient Document Retrieval with Vision Language Models
By
–
ColPali, is a novel approach for Efficient Document Retrieval using Vision Language Models developed by the team at Illuin Tech. It outperforms standard retrievers by a huge margin with less latency and complexity.
— Satya Mallick (@LearnOpenCV) 17 septembre 2024
We will explore and test this through a much demanding… pic.twitter.com/vRRECCVxEUColPali, is a novel approach for Efficient Document Retrieval using Vision Language Models developed by the team at Illuin Tech. It outperforms standard retrievers by a huge margin with less latency and complexity. We will explore and test this through a much demanding
-
OpenAI o1 breakthrough moment: advanced instruction-following for game development
By
–
When I first tried ChatGPT in 2022, it hit me: we finally have AI that could follow instructions. It was my "ChatGPT moment" – a real game-changer. Now, I've had my "o1 moment". I asked OpenAI-o1 to create a 3D snake game, then add Space Engineers elements. It did both
-
Prompting ChatGPT-4o with chain-of-thought reduces correctness
By
–
For comparison, when I give the same prompt to ChatGPT 4o, but replacing the last sentence with “Think step-by-step,” it returns 0 correct answers in 10 samples: “salubrious,” “abstemious” (x4), “illegible,” “repetitive,” “luminously”, “abstinential,” and “inexpensive.”
-
LLMs struggle with word length and final letters
By
–
It’s not always right of course — the other three answers I saw in five samples were “facetiously” (twice) and “abstentious.” But 2/5 here is pretty good — LLMs are famously bad at identifying the length or final letters of words
-
SambaNova Llama 405B faster alternative to OpenAI o1
By
–
Tired of slow thinking with OpenAI o1? Try our demo using the fastest Llama 405B on SambaNova Cloud built by @KyleLiang5
. Try it here: -
Technical analysis of ChatGPT internal message parameters
By
–
In the example conversation where ChatGPT is messaging users on its own, “default_model” is set to “auto” where normally it follows the same model name “message_type” is also empty which is normally set to “next”. AutoGPT, is it you?
-
Internal development of new AI model ‘Canmore’ currently restricted
By
–
They already do They are called “Canmore” internally but protected for now and not accessible