Warning: mkdir(): No space left on device in /var/www/group-telegram/post.php on line 37
Warning: file_put_contents(aCache/aDaily/post/datastorieslanguages/--): Failed to open stream: No such file or directory in /var/www/group-telegram/post.php on line 50 Data, Stories and Languages | Telegram Webview: datastorieslanguages/351 -
Training Large Language Models to Reason in a Continuous Latent Space
Новая статья от META - про кокосик! То есть Coconut (Chain of Continuous Thought).
Авторы предлагают изменить подход к reasoning в LLM, перемещая процесс из "language space" в "latent space". По сути, модель думает не токенами, а с использованием hidden state. Это позволяет делать breadth-first search и избегать преждевременных решений при выборе неоптимального пути. Coconut превосходит CoT в задачах логического мышления с необходимостью сложного планирования и backtracking.
Подобные идеи уже пробовали в других работах, но у META получилось довольно красиво. Кстати, в качестве базовой модели используют старый добрый GPT-2.
Training Large Language Models to Reason in a Continuous Latent Space
Новая статья от META - про кокосик! То есть Coconut (Chain of Continuous Thought).
Авторы предлагают изменить подход к reasoning в LLM, перемещая процесс из "language space" в "latent space". По сути, модель думает не токенами, а с использованием hidden state. Это позволяет делать breadth-first search и избегать преждевременных решений при выборе неоптимального пути. Coconut превосходит CoT в задачах логического мышления с необходимостью сложного планирования и backtracking.
Подобные идеи уже пробовали в других работах, но у META получилось довольно красиво. Кстати, в качестве базовой модели используют старый добрый GPT-2.
"And that set off kind of a battle royale for control of the platform that Durov eventually lost," said Nathalie Maréchal of the Washington advocacy group Ranking Digital Rights. "The argument from Telegram is, 'You should trust us because we tell you that we're trustworthy,'" Maréchal said. "It's really in the eye of the beholder whether that's something you want to buy into." But Telegram says people want to keep their chat history when they get a new phone, and they like having a data backup that will sync their chats across multiple devices. And that is why they let people choose whether they want their messages to be encrypted or not. When not turned on, though, chats are stored on Telegram's services, which are scattered throughout the world. But it has "disclosed 0 bytes of user data to third parties, including governments," Telegram states on its website. "For Telegram, accountability has always been a problem, which is why it was so popular even before the full-scale war with far-right extremists and terrorists from all over the world," she told AFP from her safe house outside the Ukrainian capital. Emerson Brooking, a disinformation expert at the Atlantic Council's Digital Forensic Research Lab, said: "Back in the Wild West period of content moderation, like 2014 or 2015, maybe they could have gotten away with it, but it stands in marked contrast with how other companies run themselves today."
from us
