Tag: LLM공부
All the articles with the tag "LLM공부".
-
The Full Flow of RAG Data Preparation — Selection, Cleansing, Chunking, Metadata, and Evaluation Sets
While working on the FEMS project, I got curious about 'what and how should go into a vector DB.' Key insight — data should be selected backward from 'questions that need to be retrieved,' not from the 'domain' criterion. Also covers cleansing / the effect of metadata (document title, section path) before chunks / OCR preprocessing for analog data / building an evaluation set — the full flow of the RAG data pipeline.
-
LangGraph Study #3 — Human-in-the-Loop (Approving Risky Tools) · Checkpoints and the Interrupt Mechanism
When an agent automatically calls risky tools like file deletion, sending emails, or payments, things go wrong. Implementing the HITL pattern using LangGraph's checkpoint + Interrupt mechanism to get human approval right before a risky tool call. An unexpected trap: the LLM's own safety guard can block execution before it ever reaches HITL. How to separate the system prompt from the HITL gate.
-
LangGraph Study Log #2 — Extending State + Dynamic System Prompts + Chatbot (and the Truth Behind 'AI Remembers')
I extended State beyond just messages to include user name, tool call count, and session start time, then used that State to dynamically build the system prompt. Along the way, building a chatbot led me to the real revelation of the day: what 'memory' actually is.
-
LangGraph Study Log #1 — My First Graph + Why State Confused Me
Starting LangGraph, one of the study candidates I had noted down. Building nodes, edges, conditional branches, and loops with StateGraph was more intuitive than expected, but the term 'State' didn't click for a while. I finally settled on it by comparing it to a game character's inventory and HP.