LLM-Resources

Websites

• Pinecone, Chunking Strategies for LLM Applications
• Ilya Rice: How I Won the Enterprise RAG Challenge
• Anthropic, Introducing Contextual Retrieval
• GPT Researcher, How we built GPT Researcher
• Reddit, Summarize this conversation in a way that can be…
  • About Memory Compression
• Langchain, Adding Long Term Memory to OpenGPTs
  • (https://github.com/langchain-ai/opengpts)[https://github.com/langchain-ai/opengpts]
  • Memory
    • Semantic Memory
    • Generative Agents
  • Questions That Determine How We Implement Memory.
    • What is the state that’s tracked?
    • How is the state updated?
    • How is the state used?
• thakkarparth007.github.io, Copilot Internals
  • https://github.com/thakkarparth007/copilot-explorer
• Github, Copilot Analysis

Tools / Frameworks

Document Processing

• PaddleOCR
• Docling
• MinerU
• PyMuPDF
• pypdfium2
  • Python binding for Pdfium
  • Used by Dify as default extractor for PDF content.
• EasyOCR
  • Used by Docling as the default OCR Engine.
• LangExtract
  • Data Extraction

Deep Research

• LangChain, Open Deep Research
• dzhng, Deep Research
• Jina AI, Deep Research
• guy-hartstein, Agentic Company Researcher
  • Internally used Tavily AI.
• GPT Researcher

LLM Papers

• Large Language Models (LLMs) on Tabular Data: Prediction, Generation, and Understanding - A Survey
  • About excel/csv cleaning
  • About Table Serialization, Row-wise Serialization, Attribute-Value Pairing
• Plan-and-Solve Prompting: Improving Zero-Shot Chain-of-Thought Reasoning by Large Language Models
  • About LLM Plan and Solve
  • https://github.com/AGI-Edgerunners/Plan-and-Solve-Prompting
  • Better CoT, especially in Math.
• Measuring and Narrowing the Compositionality Gap in Language Models
  • https://github.com/ofirpress/self-ask
  • About LLM Self-Ask
  • Solve Multi-Hop Questions
• 2022, ReAct: Synergizing Reasoning and Acting in Language Models
  • LLM ReAct
• 2023, Generative Agents: Interactive Simulacra of Human Behavior
  • About Long Term Memory, Generative Agent (Memory)
• 2023, Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection
  • Self-RAG
  • Reflection Tokens
    • Retrieve (need for retrieval)
      • Retrieve: when to retrieve X
    • Critique (generation quality)
      • IsREL (relevant): X provides useful info to solve Y
      • IsSUP (supported): all verification-worth statement in X is supported by Y
      • IsUSE (useful): X is useful response to Y
  • https://selfrag.github.io/
  • https://github.com/SauravP97/AI-Engineering-101/tree/main/self-rag
  • https://github.com/AkariAsai/self-rag
• 2024, The Prompt Report: A Systematic Survey of Prompt Engineering Techniques
  • Over 58 different types of Prompting Technique?
• 2024, Corrective Retrieval Augmented Generation
  • CRAG
    • Correct
    • Incorrect
    • Ambiguous
  • Langchain example: https://github.com/langchain-ai/langgraph/blob/main/examples/rag/langgraph_crag.ipynb?ref=blog.langchain.com
  • SauravP97 on Corrective RAG: https://github.com/SauravP97/AI-Engineering-101/tree/main/corrective-rag