Thinking models o1-3: a brief overview and what programmers can expect next
When LLMs first appeared, they were a bit like children - they said the first thing that came to their mind and didn't care much about logic. They needed to be reminded:…
Introduction to Large Language Models (LLM)
Applications and Benefits of LLM in IT
Improving RAG with Knowledge Graphs
Generation with augmented sampling is a method that connects external data sources to improve the output of large language models. This method is ideal for LLM to access…
Speculative Inference Algorithms for LLM
In recent years, the quality of LLM models has significantly improved, quantization methods have become better, and graphics cards more powerful. Nevertheless, the quality…
Basics and advanced techniques of RAG
In this post, we will talk in detail about RAG at each of its stages, its modifications, and its promising directions of development at the time of writing.
Super Ethical Reality: what to consider before using LLM in development
Should we trust something that cannot understand the consequences of its actions? The answer seems obvious, but as the niche of linguistic models develops, we increasingly…
Is there a soul in LLM?
Colleagues, don't argue and don't be afraid - I'm not going to prove the existence or absence of a soul in either robots or humans :) After all, a definition of the soul…
The most advanced LLMs make predictions about their development for 2025
I asked the following question to the most popular LLMs.
Autogeneration of tests for Java/Kotlin in IntelliJ IDEA: comparison of AI tools
For most developers, tests are the least favorite part of the job. Recently, we confirmed this by surveying more than 400 developers at the Joker and Heisenbug conferences…
How the SIBUR Innovation Laboratory is organized and why it is needed
Creating innovations quickly and flexibly can be done in a startup or a small company, but what if you are an industrial giant with more than 25 factories across the…
How LLMs are changing system architecture: from simple data pipelines to intelligent autonomous agents
At every technical conference lately, the word "agents" is sure to be mentioned. They are presented in different ways: as the next step after RAG, as a silver bullet…
Real implementation of LLM in business processes
Overview of three cases of LLM implementation in large companies: Magnit, Skyeng, and Tochka. They solve different tasks, but they all have one thing in common: competent…
How to choose the best model for coding: using SLM and local LLM
Hello, this is Yulia Rogozina, business process analyst at Sherpa Robotics. Today I have translated for you an article dedicated to the use of SLM and local LLM. Small…
ACM RecSys — 2024: trends and reports from the largest conference on ML in recommender systems
Hello! My name is Peter Zaidel and I am a senior developer at Music. Together with other guys from Yandex, who develop recommender systems in various services, I visited…
Hallucinations in LLM and methods of combating them
One of the main problems with using large language models in business is that LLMs are prone to hallucinations. How can you trust your customers to a chatbot that can…
How to improve the understanding of numbers in language models?
Hello, this is Yulia Rogozina, a business analyst at Sherpa Robotics. Today I translated for you an article about the shortcomings of language models in terms of calculations,…
Generation of augmented extraction: from theory to implementation of LangChain
From the theory of the original academic paper to its implementation in Python with OpenAI, Weaviate, and LangChain
Psst, need AI? 5 useful tools for developers
Hello, tekkix! This is Misha Stepnov, head of the R&D Big Data center at MTS Digital. If you work with language models or dream of creating your own AI assistant, it…
Forget does not mean delete: SURE is a new quantum-resistant method of "forgetting" in LLM
Hello, this is Elena Kuznetsova, an automation specialist at Sherpa Robotics. Today I translated an article for you on a very important topic, it addresses the problem…