Awesome-RAG

The sections outline is in complete draft form. Everything is in motion (in this document and in my head).

• General
  • Disadvantages of RAG
  • Patterns
• Dialogue Routing
• LLM Models
  • Finetuning and Pretraining
• Retrieval
  • Vector retrieval
    • Chunking
      • Positional chunking
      • Semantic chunking
    • Embeddings
    • Vector Search
    • RAG Fusion
  • Not Vector retrieval
• Prompts
  • Prompting strategies
    • Multi-Modal RAG
    • Multi-index RAG
    • Multi-Document
    • FLARE
    • Chain-of-Verification
  • Automated prompt optimization
  • Context
    • Long context RAG
    • Knowledge and Knowledge Graphs
• Generation
  • Hallucination
  • Guardrails
• Evaluation
• Performance and cost
• Privacy
• Security
  • Overloading context
  • Injections
• Applications of RAG
  • Chatbots
• Tools
• Vendor-specific examples
  • Elastcisearch + OpenAI
  • OpenAI and ChatGPT
    • Tools and fucntions
  • Vespa
  • Qdrant
• Running RAGs in production
• Vectors corner

General

• Retrieval Augmented Generation — Intuitively and Exhaustively Explained
• GraphRAG - Microsoft Research Blog Post

Disadvantages of RAG

• Disadvantages of RAG

RAG Patterns

• Generative AI Lifecycle Patterns
• Why do RAG pipelines fail? Advanced RAG Patterns — Part1 Ozgur Guler
• How to improve RAG peformance — Advanced RAG Patterns — Part2
• Patterns for Building LLM-based Systems & Products
• AI Engineer Summit - Building Blocks for LLM Systems & Products
• Technical Considerations for Complex RAG

Dialogue Routing

• Routing in RAG-Driven Applications

Retrieval

Vector Retrieval

• Boosting RAG: Picking the Best Embedding & Reranker models
• What We Need to Know Before Adopting a Vector Database

Chunking

• Chunking Strategies for LLM Applications
• Evaluating the Ideal Chunk Size for a RAG System using LlamaIndex
• How to Chunk Text Data — A Comparative Analysis

Positional chunking

Semantic chunking

Embeddings

Vector Search

• Awesome Search
• Advanced RAG Retrieval Strategies: Sentence Window Retrieval

RAG Fusion

• Forget RAG, the Future is RAG-Fusion

Not Vector Retrieval

• Vector Search Is Not All You Need
• Build a search engine, not a vector DB
• Improving RAG (Retrieval Augmented Generation) Answer Quality with Re-ranker
• From Search to Synthesis: Enhancing RAG with BM25 and Reciprocal Rank Fusion

Generation

Prompts

• Emerging RAG & Prompt Engineering Architectures for LLMs
• How to Cut RAG Costs by 80% Using Prompt Compression

Prompting strategies

Multi-Modal RAG

• Multi-Modal RAG

Multi-index RAG

• Having all of your data stored in one collection isn't always the best for RAG apps

Multi-Document

• Advanced RAG — Multi-Documents Agent with LlamaIndex

FLARE

• Better RAG with Active Retrieval Augmented Generation FLARE

Chain-of-Verification

• in-Of-Verification Reduces Hallucination in LLMs

Chain-Of-Thought

• Chain-Of-Thought Prompting In LLMs

Context

• The Needle In a Haystack Test
• Conversational Memory for LLMs with Langchain

Long context RAG

• The next generation of RAG: Long-Context RAG
• NVIDIA Research: RAG with Long Context LLMs

Knowledge and Knowledge Graphs

• Graph RAG: Unleashing the Power of Knowledge Graphs with LLM

• Embeddings + Knowledge Graphs: The Ultimate Tools for RAG Systems

• The Practical Benefits to Grounding an LLM in a Knowledge Graph Daniel Bukowski

• Implement RAG with Knowledge Graph and Llama-Index

• Awesome Knowledge Graphs

• HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models

Automated prompt optimization

Hallucination

• How to Detect Hallucinations in LLMs
• Measuring Hallucinations in RAG Systems

Guardrails

• Safeguarding LLMs with Guardrails
• NeMo Guardrails: The Missing Manual

LLM Models

Finetuning and Pretraining

• Fine-Tuning Llama 2.0 with Single GPU Magic
• Practitioners guide to fine-tune LLMs for domain-specific use case
• Are You Pre-training your RAG Models on Your Raw Text?
• Combine Multiple LoRA Adapters for Llama 2
• RAG vs Finetuning — Which Is the Best Tool to Boost Your LLM Application?

Evaluation of RAGs

• RAG Evaluation
• Evaluating RAG: A journey through metrics
• Exploring End-to-End Evaluation of RAG Pipelines
• Evaluation Driven Development, the Swiss Army Knife for RAG Pipelines
• Evaluating the Ideal Chunk Size for a RAG System using LlamaIndex

Performance and cost

• Secrets to Optimizing RAG LLM Apps for Better Performance, Accuracy and Lower Costs!

Privacy

• Masking PII Data in RAG Pipeline

Security

• Hijacking Chatbots: Dangerous Methods Manipulating GPTs

Applications of RAG

Chatbots

Tools

• Three Open-Source RAG Tools You Need to Know About
• HayStack
• RAGAS

DSPy

• DSPy — Does It Live Up To The Hype?

AutoRAG

• AutoRAG - AutoML tool for RAG. Automatically optimize RAG pipeline with single YAML file.

AutoGPT

Langchain

• Langchain
• Langchain is NOT for production use. Here is why ..

LlamaIndex

• LlamaIndex

• Building Production-Ready LLM Apps with LlamaIndex: Document Metadata for Higher Accuracy Retrieval

• Building Production-Ready LLM Apps With LlamaIndex: Recursive Document Agents for Dynamic Retrieval

Vendor-specific examples

• RAG Pipeline with Mistral 7B Instruct Model in Colab: A Step-by-Step Guide Qendel AI GoPenAI

Elastcisearch + OpenAI

• ChatGPT and Elasticsearch: OpenAI meets private data

OpenAI and ChatGPT

• Compare PDF Question Answering Systems Build with OpenAI and Google VertexAI

Tools and fucntions

• Unlocking the Power of the OpenAI API: Master Function-Calling with Practical Examples
• penAI/Chat-GPT Function Calling : for Enhanced AI Interactions

Vespa

• Hands-On RAG guide for personal data with Vespa and LLamaIndex

Qdrant

Running RAGs in production

Vectors corner

• Similarity Search, Part 2: Product Quantization
• Binary and Scalar Embedding Quantization for Significantly Faster & Cheaper Retrieval
• Cohere int8 & binary Embeddings - Scale Your Vector Database to Large Datasets Image of Nils Reimers

Various

• Improve RAG Pipelines With These 3 Indexing Methods

LLM Agents