Skip to content
TechFlix - Binge Read the Technology
Metrics for LLM evaluation
Type to start searching
TechFlix - Binge Read the Technology
Welcome to TechFlix
About Techflix
DSA
DSA
Top 20 DSA
Top 20 DSA
Top 20 problem DSA till dec 1
Top 20 DSA
High Level System Design
High Level System Design
High level system design index
ChatApp
ChatApp
Chat Application
Large Language Models Interview
Large Language Models Interview
Large Language Models Interview Index
01 How to navigate the course
01 How to navigate the course
Navigate the course Free Preview
02 Prompt Engineering basics of LLM
02 Prompt Engineering basics of LLM
What is the difference between Predictive/Discriminative AI and Generative AI? Free Preview
What is LLM & how LLMs are trained? Free Preview
What is a token in language model?
How to estimate the cost of running a SaaS-based & Open source LLM model?
Explain Temperature parameter and how to set it?
What are different decoding strategies for picking output tokens?
What are different ways to define stopping criteria in large language model?
How to use stop sequences in LLMs?
Explain the basic structure of prompt engineering?
Explain In-Context learning?
Types of prompt engineering
Key considerations for few-shot prompting
Strategies for writing good prompts
What is hallucination & how can it be controlled using prompt engineering?
Improving LLM reasoning ability through prompt engineering
How to improve LLM reasoning if your COT prompt fails?
Test yourself
03 Retrieval augmented generation RAG Systems
03 Retrieval augmented generation RAG Systems
Increasing accuracy, reliability & making answers verifiable in LLM
How does Retrieval augmented generation (RAG) work?
Benefits of using RAG system
Architecture patterns for customizing LLM with proprietary data
When to use Fine-tuning instead of RAG
Test yourself
04 Data Chunking
04 Data Chunking
What is chunking and why do we chunk our data?
Factors influencing chunk size
Different types of chunking methods
Finding the ideal chunk size
Test yourself
05 Embedding Model
05 Embedding Model
What are vector embeddings? And what is an embedding model?
Using embedding models in LLM applications
Differences in embedding short and long content
Benchmarking embedding models on your own data
Scenario-based question
Improving sentence transformer models for embedding
Test yourself
06 Vector DB
06 Vector DB
What is a vector DB?
Differences between vector DB and traditional databases
How a vector database works
Vector index, vector DB & vector plugins differences
Different vector search strategies
Scenario-based question
How clustering reduces search space and mitigates failures
Random projection index
Locality-sensitive hashing (LHS) indexing method
Product quantization (PQ) indexing method
Comparing vector indexes for specific scenarios
Ideal search similarity metrics for use cases
Types and challenges in vector DB filtering
Determining the best vector database for your needs
Test yourself
07 Search
07 Search
Importance of effective search
Information retrieval & semantic search architecture patterns
Achieving efficient, accurate search in large datasets
Scenario-based question
Keyword-based retrieval method
Fine-tuning re-ranking models
Common information retrieval metrics and their limitations
Scenario-based question
Metrics for evaluating recommendation systems
Comparing information retrieval metrics
How hybrid search works
Scenario-based question
Handling multi-hop/multifaceted queries
Techniques to improve retrieval
Test yourself
08 Large Language Models
08 Large Language Models
Detailed explanation of self-attention
Disadvantages of self-attention and solutions
What is positional encoding?
Detailed Transformer architecture explanation
Advantages of transformers over LSTM
Differences between local and global attention
Overcoming computational and memory demands of transformers
Increasing context length of LLMs
Optimizing transformer architecture for large vocabularies
Scenario-based question
Types of LLM architectures and best use cases
Test yourself
09 Supervise Fine tuning SFT LLM
09 Supervise Fine tuning SFT LLM
What is fine-tuning and why it's needed in LLM?
When to fine-tune LLM
Deciding on fine-tuning
Creating fine-tuning datasets for Q&A
Improving model responses based on context sufficiency
Setting hyperparameters for fine-tuning
Estimating infrastructure requirements for fine-tuning LLM
Fine-tuning LLM on consumer hardware
Categories of the PEFT method
Understanding catastrophic forgetting in LLMs
Re-parameterized methods for fine-tuning LLM
Test yourself
10 Deployment
10 Deployment
Why quantization doesn't decrease LLM accuracy
Techniques for optimizing LLM inference for higher throughput
Test yourself
11 Hallucination
11 Hallucination
Different forms of hallucinations
Controlling hallucinations at various levels
Test yourself
12 Evaluation
12 Evaluation
Evaluating the best LLM model for use cases
Evaluating RAG-based systems
Metrics for LLM evaluation
Explaining the Chain of verification
Test yourself
13 Agents
13 Agents
Basic concepts of agents and implementation strategies
Need for agents and common implementation strategies
ReAct prompting with code example
Plan and Execute prompting strategy
OpenAI functions strategy with code examples
OpenAI functions vs LangChain Agents
Test yourself
14 Prompt Hacking
14 Prompt Hacking
What is prompt hacking and its importance?
Types of prompt hacking
Defense tactics against prompt hacking
Test yourself
15 Preference Alignment in LLMs
15 Preference Alignment in LLMs
Deciding on Preference alignment over SFT
What is RLHF and its use
Understanding reward hacking
Different preference alignment methods
16 Miscellaneous
16 Miscellaneous
Optimizing LLM system cost
What are Mixture of Expert models (MoE)?
Building a production-grade RAG system in detail
Understanding FP8 variable and its advantages
Training LLM with low precision without compromising accuracy
Calculating size of KV cache
Dimensions of each layer in multi-headed transformation attention block
Ensuring attention layer focuses on relevant input parts
17 Case Studies
17 Case Studies
Case Study 1: LLM Chat Assistant with dynamic context
Case Study 2: Prompting Techniques
Low Level System Design
Low Level System Design
Low Level System Design Index
Chess - Low Level System Design
Splitwise - Low Level System Design
Consolidate LLD
Consolidate LLD
1.ChessLLD
MTech
MTech
MTech Subjects
Semester 1
Semester 1
AI
AI
Problem Solving Agent Using Search
MFML
MFML
Bayes Theorem Example
Bayes basic
Identifying Components of Bayes' Theorem
Matrices
SVD
System of linear equation
Vector Norms
Vector space
Math Foundation
Math Foundation
Syllabus
Vectors Matrices
Systems of Linear Equations and Their Solutions
Eigenvalues and Eigenvectors
Multivariate Calculus
4 Multivatiate calculus
Gradient Descent
tochastic Gradient Descent (SGD)
Optimization for Support Vector Machines (SVM)
Maths
Maths
ExamPapers
Previous Year paper
Python for MLAI
Python for MLAI
ClassNotes
ClassNotes
Session 1
Session 1
Session 1
Session 2
Session 2
Session 2
Session 3
Session 3
Session 3
Session 3.1 : Functions in Python
Statistics
Statistics
1 Statistics
3 mean
Lecture1
MachineLearningRoadmap
MachineLearningRoadmap
MachineLearningRoadmap
Projects
Projects
Generative-System-Design
Metrics for LLM evaluation
#
Content for Metrics for LLM evaluation goes here.