Generative AI and LLM Mesh

• Introduction
  • Dataiku Generative AI Capabilities
  • Using the LLM Mesh
• LLM connections
  • Hosted LLM APIs
  • Locally-running HuggingFace models
• Running Hugging Face models
  • Setup & prerequisites
  • Text generation
  • Text embedding
  • Text reranking
  • Other model types
  • Monitoring model instances
  • The model cache
• Prompt Studios
• Prompt Recipe
  • When to use a Prompt Recipe
  • Prompt Recipes in the LLM Mesh ecosystem
• Adding Knowledge to LLMs
  • Introduction to Knowledge Banks and RAG
  • Initial setup
  • Working with Vector stores
  • Embedding & searching datasets
  • Embedding and searching documents
  • Search settings
  • RAG guardrails
  • Document-Level Security
  • GraphRAG
  • Automated RAG Optimization
• Guardrails
  • Introduction to guardrails
  • PII detection
  • Prompt injection detection
  • Toxicity detection
  • Topics Boundaries
  • Bias Detection
  • Custom Guardrails
• LLM Mesh API
  • OpenAI-compatible API
  • Persisted conversations
• Multimodal capabilities
  • Image input
  • Image output
• Fine-tuning
  • Setup
  • Using the Fine-tuning recipe
  • Using Python code
• Cost Control
  • Concepts
  • Limitations
  • Permissions and Visibility
  • Setup (Admin only)
• Rate Limiting
  • Concepts
  • Rule Configuration
  • Limitations
  • Setup
• Evaluating LLMs & GenAI use cases
  • Overview
  • Recipe configuration
• Automated Prompt Optimization
  • Inputs
  • Output
  • Parameters