0
1

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

Building LLM Powered Applications AI(20)

Last updated at Posted at 2024-11-16

Building LLM Powered Applications
https://learning.oreilly.com/library/view/building-llm-powered/9781835462317/

References

1 Introduction to Large Language Models

Attention is all you need: 1706.03762.pdf (arxiv.org)
Possible End of Humanity from AI? Geoffrey Hinton at MIT Technology Review’s EmTech Digital: https://www.youtube.com/watch?v=sitHS6UDMJc&t=594s&ab_channel=JosephRaczynski
The Glue Benchmark: https://gluebenchmark.com/
TruthfulQA: https://paperswithcode.com/dataset/truthfulqa
Hugging Face Open LLM Leaderboard: https://huggingface.co/spaces/optimum/llm-perf-leaderboard
Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge: https://arxiv.org/abs/1803.05457

2 LLMs for AI-Powered Applications

LangChain repository: https://github.com/langchain-ai/langchain
Semantic Kernel documentation: https://learn.microsoft.com/en-us/semantic-kernel/get-started/supported-languages
Copilot stack: https://build.microsoft.com/en-US/sessions/bb8f9d99-0c47-404f-8212-a85fffd3a59d?source=/speakers/ef864919-5fd1-4215-b611-61035a19db6b
The Copilot system: https://www.youtube.com/watch?v=E5g20qmeKpg

3 Choosing an LLM for Your Application

GPT-4 Technical Report. https://cdn.openai.com/papers/gpt-4.pdf
Train short, test long: attention with linear biases enables input length extrapolation. https://arxiv.org/pdf/2108.12409.pdf
Constitutional AI: Harmlessness from AI Feedback. https://arxiv.org/abs/2212.08073
Hugging Face Inference Endpoint. https://huggingface.co/docs/inference-endpoints/index
Hugging Face Inference Endpoint Pricing. https://huggingface.co/docs/inference-endpoints/pricing
Model Card for BioMedLM 2.7B. https://huggingface.co/stanford-crfm/BioMedLM
PaLM 2 Technical Report. https://ai.google/static/documents/palm2techreport.pdf
Solving Quantitative Reasoning Problems with Language Models. https://arxiv.org/abs/2206.14858
Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena. https://arxiv.org/abs/2306.05685

4 Prompt Engineering

ReAct approach: https://arxiv.org/abs/2210.03629
What is prompt engineering?: https://www.mckinsey.com/featured-insights/mckinsey-explainers/what-is-prompt-engineering
Prompt engineering techniques: https://blog.mrsharm.com/prompt-engineering-guide/
Prompt engineering principles: https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/advanced-prompt-engineering?pivots=programming-language-chat-completions
Recency bias: https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/advanced-prompt-engineering?pivots=programming-language-chat-completions#repeat-instructions-at-the-end
Large Language Model Prompt Engineering for Complex Summarization: https://devblogs.microsoft.com/ise/2023/06/27/gpt-summary-prompt-engineering/
Language Models are Few-Shot Learners: https://arxiv.org/pdf/2005.14165.pdf
IMDb dataset: https://www.kaggle.com/datasets/yasserh/imdb-movie-ratings-sentiment-analysis/code
ReAct: https://arxiv.org/abs/2210.03629
Chain of Thought Prompting Elicits Reasoning in Large Language Models: https://arxiv.org/abs/2201.11903

5 Embedding LLMs within Your Applications

LangChain’s integration with OpenAI – https://python.langchain.com/docs/integrations/llms/openai
LangChain’s prompt templates – https://python.langchain.com/docs/modules/model_io/prompts/prompt_templates/
LangChain’s vector stores – https://python.langchain.com/docs/integrations/vectorstores/
FAISS index – https://faiss.ai/
LangChain’s chains – https://python.langchain.com/docs/modules/chains/
ReAct approach – https://arxiv.org/abs/2210.03629
LangChain’s agents – https://python.langchain.com/docs/modules/agents/agent_types/
Hugging Face documentation – https://huggingface.co/docs
LangChain Expression Language (LCEL) – https://python.langchain.com/docs/expression_language/
LangChain stable version – https://blog.langchain.dev/langchain-v0-1-0/

6 Building Conversational Applications

Example of a context-aware chatbot. https://github.com/shashankdeshpande/langchain-chatbot/blob/master/pages/2_%E2%AD%90_context_aware_chatbot.py
Knowledge base for the AI travel assistant. https://www.minube.net/guides/italy
LangChain repository. https://github.com/langchain-ai

7 Search and Recommendation Engines with LLMs

Recommendation as Language Processing (RLP): A Unified Pretrain, Personalized Prompt & Predict Paradigm (P5). https://arxiv.org/abs/2203.13366
LangChain’s blog about featurestores. https://blog.langchain.dev/feature-stores-and-llms/
Feast. https://docs.feast.dev/
Tecton. https://www.tecton.ai/
FeatureForm. https://www.featureform.com/
Azure Machine Learning feature store. https://learn.microsoft.com/en-us/azure/machine-learning/concept-what-is-managed-feature-store?view=azureml-api-2

8 Using LLMs with Structured Data

Chinook Database: https://github.com/lerocha/chinook-database/tree/master/ChinookDatabase/DataSources
LangChain File system tool: https://python.langchain.com/docs/integrations/tools/filesystem
LangChain Python REPL tool: https://python.langchain.com/docs/integrations/toolkits/python

9 Working with Code

The open-source version of the Code Interpreter API: https://github.com/shroominic/codeinterpreter-api
StarCoder: https://huggingface.co/blog/starcoder
The LangChain agent for the Python REPL: https://python.langchain.com/docs/integrations/toolkits/python
A LangChain blog about the Code Interpreter API: https://blog.langchain.dev/code-interpreter-api/
The Titanic dataset: https://www.kaggle.com/datasets/brendan45774/test-file
The HF Inference Endpoint: https://huggingface.co/docs/inference-endpoints/index
The CodeLlama model card: https://huggingface.co/codellama/CodeLlama-7b-hf
Code Llama: Open Foundation Models for Code, Rozière. B., et al (2023): https://arxiv.org/abs/2308.12950
The Falcon LLM model card: https://huggingface.co/tiiuae/falcon-7b-instruct
The StarCoder model card: https://huggingface.co/bigcode/starcoder

10 Building Multimodal Applications with LLMs

Source code for YouTube tools: https://github.com/venuv/langchain_yt_tools
LangChain YouTube tool: https://python.langchain.com/docs/integrations/tools/youtube
LangChain AzureCognitiveServicesToolkit: https://python.langchain.com/docs/integrations/toolkits/azure_cognitive_services

11 Fine-Tuning Large Language Models

Training dataset: https://huggingface.co/datasets/imdb
HF AutoTrain: https://huggingface.co/docs/autotrain/index
BERT paper: Jacob Devlin, Ming-Wei Chang, Kenton Lee, Kristina Toutanova, 2019, BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding: https://arxiv.org/abs/1810.04805

12 Responsible AI

Reducing Gender Bias Amplification using Corpus-level Constraints: https://browse.arxiv.org/pdf/1707.09457.pdf
ChatGPT racist and sexist outputs: https://twitter.com/spiantado/status/1599462375887114240
GitHub repository for an aligned dataset: https://github.com/Zjh-819/LLMDataHub#general-open-access-datasets-for-alignment-
AI Act: https://www.europarl.europa.eu/RegData/etudes/BRIE/2021/698792/EPRS_BRI(2021)698792_EN.pdf
Prompt hijacking: https://arxiv.org/pdf/2211.09527.pdf
AI Act: https://www.europarl.europa.eu/news/en/headlines/society/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence
Blueprint for an AI Bill of Rights: https://www.whitehouse.gov/ostp/ai-bill-of-rights/

13 Emerging Trends and Innovations

GPT-4V(ision) System Card: GPTV_System_Card.pdf (openai.com)
AutoGen paper: Qingyun Wu et al., 2023, AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation: https://arxiv.org/pdf/2308.08155.pdf
AutoGen GitHub: https://github.com/microsoft/autogen/blob/main/notebook/agentchat_web_info.ipynb
DALL-E 3: James Betker, Improving Image Generation with Better Captions: https://cdn.openai.com/papers/dall-e-3.pdf
Notion AI: https://www.notion.so/product/ai
Coca-Cola and Bain partnership: https://www.coca-colacompany.com/media-center/coca-cola-invites-digital-artists-to-create-real-magic-using-new-ai-platform
Malbek and ChatGPT: https://www.malbek.io/news/chat-gpt-malbek-unveils-generative-ai-functionality
Microsoft Copilot: https://www.microsoft.com/en-us/microsoft-365/blog/2023/09/21/announcing-microsoft-365-copilot-general-availability-and-microsoft-365-chat/

0
1
0

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
0
1

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?