We're in Seattle for #CVPR2024 this week!

Excited to see the #ChatDev team pushing the boundaries of LLM-powered multi-agent collaboration with their curated collection of seminal papers. Dive into the latest advancements and explore the interactive e-book here: thinkwee.top/multiagent_ebo… 📚🤖 #AI #Research #Innovation

🚀 Exciting day at #Tsinghua! The 2024 Summer School for Large Language Models kicks off today. 🌐 CS students from around the globe have gathered here, eager to dive into the latest IT trends and embark on this journey of knowledge and innovation. 📚

Introducing Internet of Agents (IoA) - a novel framework for AI agent collaboration! 🌐🤖 Imagine a world where heterogeneous AI agents located on different devices work together seamlessly through the Internet, just like humans do on the Internet. That's IoA!

🏆 We're thrilled to announce that our paper "Scaling Laws For Dense Retrieval"won the SIGIR'24 Best Paper Award! Congratulations to our research team for information retrieval THUIR from #DCST, #Tsinghua! 📚 #SIGIR #SIGIR2024 dl.acm.org/doi/abs/10.114…

1/5 🚀 Excited to share our latest paper on Configurable Foundation Models! 🧠 Inspired by the human brain's functional specialization, we propose a concept: Configurable Foundation Model, a modular approach to LLMs.

Optima explores the evolution of agent communication and scaling laws. Intriguing findings from Weize Chen —join the conversation! 💬 #TechTalk #AI #THUNLP

🚀 Excited to share our latest work: “RAGEval”! 🎉 It’s a versatile framework for generating scenario-specific RAG evaluation datasets, complete with comprehensive metrics. Perfect for rapid evaluations! 🔍 ✨ Paper: arxiv.org/abs/2408.01262

(Repost) We are thrilled to introduce our new work 🔥#SparsingLaw🔥, a comprehensive study on the quantitative scaling properties and influential factors of the activation sparsity within LLMs.💪 📎Arxiv: arxiv.org/pdf/2411.02335 📎Codes: github.com/thunlp/Sparsin… 🧵1

1/4 🚀 Densing Law of LLMs 🚀 OpenAI's Scaling Law showed how model capabilities scale with size. But what about the trend toward efficient models? 🤔 We introduce "capacity density" and found an exciting empirical law: LLMs' capacity density grows EXPONENTIALLY over time!