LLM
A Large Language Model (LLM) is a natural language processing model built on deep learning technology and trained on large-scale text data. Core capabilities include text understanding, generation, reasoning, and dialogue, enabling it to handle diverse language tasks such as translation, summarization, code generation, and question answering. LLMs typically adopt the Transformer architecture, with parameter counts ranging from billions to hundreds of billions, and serve as a foundational component of contemporary AI applications.
Core Facts
Mainstream Agent architectures typically use a large language model as the brain, combined with memory modules, planning modules, and tool interfaces to accomplish complex tasks
The tool-calling (Function Calling) mechanism allows the model to invoke external tools during reasoning, such as file read/write operations, terminal commands, and search APIs.
Mainstream approaches to agent evaluation include LLM-as-Judge, trajectory matching, and outcome verification.
Timeline (last 90 days)
Directions such as LLM, Agentic, VLA, multimodal, robotics, or foundation models have become the absolute focus of industry ML research since 2023, with many related internship positions and high salaries, but fierce competition.
Each routing decision by the Supervisor is an independent LLM call that needs to feed the complete history of messages back into the model as input, so the same stretch of history is billed repeatedly.
The ability needed to define a problem depends on a deep understanding of business context, team capabilities, the implicit rules of legacy systems, and needs users have not yet expressed, and this information is rarely fully recorded in public texts, so LLMs cannot obtain it directly from training data.
The typical problem with LangChain-based LLM Agents is that every round of decision-making requires a call to a large model, so latency and token consumption grow linearly with the number of conversation turns.
Markdown is currently the most suitable intermediate format for LLMs and RAG to process PDFs, because it preserves structures such as headings, lists, and tables while being far smaller in size than the original PDF.
A token is the basic unit of text processed by an LLM, and it can be a complete word, part of a word, punctuation, or a space.
When an LLM generates text, it first computes a probability distribution over the entire vocabulary (usually tens of thousands to over a hundred thousand tokens) for each token, and then samples the next token according to the probabilities.
In AI agent development, n8n typically plays the role of the execution layer: receiving the LLM's decision results, calling APIs, reading and writing data, sending notifications, or handing off to a human.
Traditionally, running and testing LLMs required powerful GPU clusters and expensive cloud services, but with Docker Desktop and open-source models, an ordinary laptop can become an AI security experimentation platform.
When handling form structures, LLMs often treat the entire HTML or JSON block as field content to be returned wholesale, lacking the semantic judgment that "this is an example, not a real value."
40 more timeline events
All Facts (20)
Mainstream Agent architectures typically use a large language model as the brain, combined with memory modules, planning modules, and tool interfaces to accomplish complex tasks
90%VerifiedThe tool-calling (Function Calling) mechanism allows the model to invoke external tools during reasoning, such as file read/write operations, terminal commands, and search APIs.
90%VerifiedMainstream approaches to agent evaluation include LLM-as-Judge, trajectory matching, and outcome verification.
90%VerifiedGenerative AI refers to artificial intelligence systems that can automatically generate text, images, code, and other content based on input prompts.
90%VerifiedPrompt注入（Prompt Injection）是针对大语言模型应用的安全威胁，攻击者通过在输入中嵌入恶意指令试图覆盖系统原有行为约束
90%VerifiedThe root cause of AI hallucination lies in the generative mechanism of LLMs: the model is essentially performing next-token probability prediction, optimizing for linguistic coherence rather than factual accuracy.
90%VerifiedLLMs exhibit a 'Lost in the Middle' phenomenon, where the model pays more attention to information at the beginning and end of the context while tending to overlook information in the middle.
90%VerifiedAI programming assistance tools such as GitHub Copilot, Cursor, and Claude are essentially code completion and generation systems formed by training large language models on massive code repositories.
90%VerifiedAI Agent通过工具调用（Tool Use）机制与外部系统交互，包括执行代码、调用API、读写数据库等
90%VerifiedRAG（检索增强生成）技术架构将文档切片、向量化、存入向量数据库，查询时先检索相关片段再交给模型生成回答
90%VerifiedThe platform adopts a design philosophy of AI generation with human editability, where AI handles the first draft and the final judgment and revision authority is handed over to testers.
85%VerifiedThe output of LLMs is inherently non-deterministic; even with the same Prompt as input, the model may return results with different formats, content, or even logic at different times.
85%VerifiedAI evaluation targets are divided into three layers—large language model evaluation, knowledge base (RAG) evaluation, and agent (Agent) evaluation—forming a progressive relationship of increasing complexity and knowledge depth.
80%VerifiedLLM output is non-deterministic; due to the non-associativity of floating-point operations and differences in context windows, even with identical inputs, outputs are difficult to reproduce exactly and do not possess idempotency.
80%VerifiedThe Function Calling mechanism allows LLMs to declare in structured JSON format the external tools and parameters they need to invoke, and the tool's return value is injected as-is into the next round of model input.
80%VerifiedRAG effectively mitigates the two major problems of large language models—knowledge cutoff and hallucination—by forcing the model to extract answers from trusted documents.
80%VerifiedLarge language models are essentially stateless functions whose weights are fixed at inference time and do not learn over time.
80%VerifiedThe emergent capability of large language models refers to the complex reasoning ability that suddenly manifests after the parameter scale surpasses a certain threshold.
80%VerifiedThe LLM tool supports multiple model providers including OpenAI, Anthropic, and Google.
80%VerifiedRAG（检索增强生成）的核心思想是在大模型生成回答之前，先从外部知识库中检索相关文档片段作为上下文注入Prompt
80%