Scribd
Upload
9 views12 pages

Ibm

Large Language Models (LLMs) are programs that recognize, summarize, translate, and predict human language using sequential text data and extensive training on large datasets. Key processes in creating LLMs include data collection, pre-processing, model architecture (typically transformer-based), training, evaluation, and deployment. LLMs have various applications in content creation, business productivity, education, and accessibility, and techniques like Retrieval Augmented Generation (RAG) and prompt engineering enhance their performance and accuracy.

Uploaded by

Sai tejaswini Nagella
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
9 views12 pages

Ibm

Large Language Models (LLMs) are programs that recognize, summarize, translate, and predict human language using sequential text data and extensive training on large datasets. Key processes in creating LLMs include data collection, pre-processing, model architecture (typically transformer-based), training, evaluation, and deployment. LLMs have various applications in content creation, business productivity, education, and accessibility, and techniques like Retrieval Augmented Generation (RAG) and prompt engineering enhance their performance and accuracy.

Uploaded by

Sai tejaswini Nagella
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 12

EVERYTHING

ABOUT
LLM
LARGE LANGUAGE MODEL(LLM)
An program that can recognize ,
summarize, translate, predict human
language(text).

Works on sequential text data.

Trained on large amounts of text data .


Hence got the name LLM.

This training enables it to learn complex


patterns and relationships within human
language.

The neural network architecture that is


HOW LLM’s ARE CREATED
Data collection – Very large raw data.

Pre-processing – Clean ,Tokenize, etc…

Model architecture – Transformer.

Training – Training on preprocessed


data.

Evaluation – Accuracy , Human


evaluation.

Deployment
SOME FAMOUS LLM’s
GPT-4 (Generative Pre-trained
Transformer)
• By OpenAI in 2023.
• Trained on over 1.7 trillion
parameters.

BERT (Bidirectional Encoder


Representations from Transformers)
• By Google in 2018.
• Trained on over 300Million
parameters.

LaMDA(Language Model for Dialogue


Applications)
APPLICATIONS OF LLM’s
Content Creation
• Automated writing ,Creative work,
Code generation.

Business & Productivity.


• Costumer support, Data analysis
etc..

Education & Research


• Personalized learning, Academic
research , etc..

Accessibility and Inclusion


• Speech to text , translation, word
A TECHNIQUE USED IN LLM’s
Retrieval Augmented
Generation(RAG):

 RAG combines the power of LLMs with


the ability to retrieve relevant
information from external sources
(like databases, documents, or APIs).

 Ex:
Imagine an LLM-powered chatbot
for a customer service department.
Without RAG, the chatbot might provide
generic or inaccurate responses. With
RAG, the chatbot can access the
A TECHNIQUE USED IN LLM’s
Retrieval Augmented
Generation(RAG):
UNLOCKING POTENTIAL OF LLM’s
Prompt Engineering :

 Prompt engineering is crafting input


queries or instructions to get more
accurate outputs from large language
models (LLMs).

 Prompt engineering is the key to


unlocking full potential of a LLM.
 Guides the LLM to give desired
output.
 can help minimize the risk of
generating harmful, biased, or
nonsensical output.
UNLOCKING POTENTIAL OF LLM’s
Prompt Engineering Methods:

 Zero-Shot Learning
 Giving LLM a direct instruction or
question without providing any
examples.

 One-Shot Learning
 Provides a single example for
guidance.

 Few-Shot Learning
 Provides multiple examples for
improved learning and
performance.
USING LLM IN CUSTOM APPLICATIONS
API & LangChain:

 Application Programming
Interface(API)-
• An API is a set of rules and
protocols that allows different
software applications to
communicate and exchange
information with each other.

 LangChain-
• LangChain is a powerful open-
source framework designed to
develop applications powered by
large language models (LLMs).
USING LLM IN CUSTOM APPLICATIONS
LangChain:
OpenAI
APPLICATION Gemini
Claude
U
LangC Cohere
hain
Jurassic
LLAMA
USE Fronte Backe
R nd nd Utilities Mistral
Proplexity
Thank you
U

You might also like