#logicaldatamodel kết quả tìm kiếm

Today Thinking Machines Lab is launching our research blog, Connectionism. Our first blog post is “Defeating Nondeterminism in LLM Inference” We believe that science is better when shared. Connectionism will cover topics as varied as our research is: from kernel numerics to…

The standard way to improve reasoning in LLMs is to train on long chains of thought. But these traces are often brute-force and shallow. Introducing RLAD, where models instead learn _reasoning abstractions_: concise textual strategies that guide structured exploration. 1/N🧵

Is your law firm drowning in paperwork? 📄➡️💻 Discover how to streamline your workflows, safeguard sensitive data, and boost productivity with our legal document management solution. Say goodbye to scattered files and hello to efficient, secure, and compliant case management.…

📢 Releasing our latest paper For LLMs doing reasoning, we found a way to save up to 50% tokens without impacting accuracy It turns out that LLMs know when they’re right and we can use that fact to stop generations early WITHOUT impacting accuracy.

1 /n The Language of Thought: When AI Speaks Prolog, Things Get Interesting Large Language Models (LLMs) have undeniably revolutionized how we interact with and leverage the power of language. They can generate human-quality text, translate languages, and even write different…

RAGで意外と悩むのが、「用意したこのデータ、LLMにとって本当に必要なのか？」という点です。 実際、既に知っているような話を追加で聞かされるとモデルの性能が落ちることが分かっています。 そんなときは「まずはLLMにどの情報が役立つか判断させる」流れが有効だと示唆されています。…

A big part of our mission at Thinking Machines is to improve people’s scientific understanding of AI and work with the broader research community. Introducing Connectionism today to share some of our scientific insights.

What does it take to monitor 4.5M+ devices and trillions of metrics daily? LogicMonitor CTO Nitin Navare breaks down real-world #observability, agentic #AIOps & why we actually use our own product to run our platform: ow.ly/Uwv350Xkzlx

Dear friends, I’m so excited to introduce Living Encyclopedia: Large Nature Model (LNM)—an AI-powered platform of DATALAND that reimagines the traditional concept of an encyclopedia by transforming vast nature-based datasets into immersive, multi-sensory experiences.

In the year since LRMs ("reasoning models") hit the scene, we have been trying to understand, analyze and demystify them.. Here are our efforts to date--conveniently all in one place (First..) Evaluation of LRMs on Planning 📜arxiv.org/abs/2409.13373 (9/24)&…

From Model Centric to Data Centric AI: A New Paradigm for Reliable Intelligent Systems in Industry

Logistic regression is one of the simplest models in machine learning, and one of the most revealing. It shows how to move from geometric intuition to probabilistic reasoning. Mastering it sets the foundation for everything else. Let’s dissect it step by step!

Advice for AI engineers 💡 A small Visual Language Model fine-tuned on your custom dataset is as accurate as GPT-5... ... and costs 50 times less. For example, LFM2-VL-3B by @LiquidAI_ ↓

Extending an LLM for new knowledge sources is tedious—fine-tuning is expensive/causes forgetting, LoRA is restrictive. Excited to share our work where we show that an LLM can be efficiently *composed* with specialized (L)LMs to enable new tasks! arxiv.org/abs/2401.02412 🧵(1/8)

NAIST、理研、京都工繊大、東京科学大学の研究者らは、LLMに”言葉の裏の意味”（皮肉など）を分からせる手法を発表しています。 『グライスの「協調の原理」や「関連性理論」』という考え方をLLMへ簡単に与えるだけで、能力が明確に向上するそうです。 これらの名称をプロンプトを入力するだけでも…

How do LLMs learn to reason from data? Are they ~retrieving the answers from parametric knowledge🦜? In our new preprint, we look at the pretraining data and find evidence against this: Procedural knowledge in pretraining drives LLM reasoning ⚙️🔢 🧵⬇️

When looking to clarify a situation, make a decision or plan your next milestone, use the Logical Levels as a powerful visualization tool. Identify the level where you are, then work your way up then down until you reach a full understanding of what you have - and what you need.

honestly, this is the best visualization of how large language models work that i've seen for a long time. > 3d interactive app, plays like a game > follow transformer operations step-by-step > the real architecture of Llama 8B > click on any layer, see formula and explanation…

Deconstructing RAG It can be hard to follow all of the RAG strategies that have come out over the past months. I created a few guides to organize them into major themes and show how to build multi-modal / semi-structured RAG on complex docs (w/ images, tables). Here's a few…

We've released RiddleBench, a new generative reasoning benchmark to truly test LLMs, based on logical reasoning questions. Current benchmarks don't capture the flexible, multi-step reasoning that is a cornerstone of human intelligence. RiddleBench aims to close that gap.