When building a Telco LLM (Large Language Model), effectively processing text data is paramount. A crucial component in this process is the 'tokenizer.' Tokenizers split text into smaller units that the model can understand, and their performance can vary significantly depending on the language. Optimized tokenizers can greatly enhance model performance. This article will delve into the basic concept of tokenizers, their importance, and how SKT’s optimized tokenizer holds a competitive edge.

As we promote projects related to Large Language Models (LLMs) with various departments in the company, we encounter individuals who either have excessive expectations about the LLMs' inference capabilities and potential, or conversely, have a critical perspective due to issues like hallucination and accuracy limitations. While LLMs are powerful and useful tools, they are not a silver bullet that can solve all problems.

When choosing and using Large Language Models (LLMs) for business, it is important they fit well with your services. If they are hard to use, they won't be very helpful. SK telecom is not only enhancing LLMs from different providers but also creating its own models called A.X LLM. Before, developers had to use different endpoints and API keys to access these models, which made things complicated. To make it simpler, we created the Global Intelligence Platform (GIP), offering all these models through one easy-to-use endpoint.

Customer service calls are a critical touchpoint in the telecom industry, and a significant amount of time spent on post-call work. During the calls, agents greet customers, assess their needs, solve problems, and retrieve relevant information in real-time. There is a large burden on customer service agents to understand customer intent and search knowledge databases. After the call, agents perform important follow-up activities like sending information to the customer, summarizing the call, and reporting to supervisors. Additionally, due to the constant influx of calls, agents sometimes receive calls while handling post-call tasks, adding to the cognitive burden.

Benchmarking Large Language Models (LLMs) is a tricky endeavor. There has been a lot of focus on model performance, but results on standard benchmark suites tend to have very little correlation with real-world use cases. Additionally, aside from performance, there are other factors (especially cost and latency) that need to be deeply considered when applying and adopting LLMs to real-world use cases, services, and products. To paint a more holistic picture on metrics that are truly important to business cases, we have created the LLM Capacity Index. This Index measures not only LLM performance, but also weighs model speed, cost, efficiency, and time to market.

Benchmarking the performance of the Telco LLM that is tuned with well-designed data is a key element in the development of the Telco LLM. SKT's AI Tech Collaboration Group comprehensively measures the performance of LLMs on a battery of tasks, ranging from general tasks that measure the reasoning ability or language ability of general models to telco-specific tasks that measure the ability to perform tasks specialized to the telco domain.

Data is the life-blood of any LLM, so in order to create a highly-performant Telco LLM, it is important to have intimate knowledge of the Telco domain and create relevant data to learn Telco business logic. Just as it takes the finest ingredients to craft a delicious meal, data is the key ingredient to building and optimizing the Telco LLM. While data is the key ingredient, a well-designed “recipe” makes or breaks the dish.

We’ve created a foundation model for the telecommunications domain we’ve deemed the Telco LLM. This large language model is the culmination of our accumulated experience in developing, tuning, and deploying large language models. The Telco LLM is the latest milestone in advancing AI technology and utilizing it in customer-facing products and services.