chapter ten

10 Large Language Models in the real world

 

This chapter covers

  • Understanding how conversational LLMs like ChatGPT work
  • Jailbreaking an LLM to get it to say things its programmers don’t want it to say
  • Recognizing errors, misinformation, and biases in LLM output
  • Fine-tuning LLMs on your own data
  • Finding meaningful search results for your queries (semantic search)
  • Speeding up your vector search with Approximate Nearest Neighbor Algorithm
  • Generating fact-based well-formed text with LLMs

10.1 Large Language Models (LLMs)

10.1.1 Scaling up

10.1.2 Guardrails (filters)

10.1.3 Red teaming

10.1.4 Smarter, smaller LLMs

10.1.5 Generating warm words using the LLM temperature paramater

10.1.6 Creating your own Generative LLM

10.1.7 Fine-tuning your generative model

10.1.8 Nonsense (hallucination)

10.2.4 Choose your index

10.2.5 Quantizing the math

10.2.6 Pulling it all together with haystack

10.2.7 Getting real

10.2.8 A haystack of knowledge

10.2.9 Answering questions

10.2.10 Combining semantic search with text generation

10.2.11 Deploying your app in the cloud

10.2.12 Wikipedia for the ambitious reader

10.2.13 Serve your "users" better

10.3 Test yourself