Mendalami Model Bahasa Besar
Kursus ini menyajikan pengantar yang komprehensif dan mendalam mengenai sejarah pengembangan model bahasa besar (LLM), arsitektur teknis inti mereka, paradigma pelatihan (pretraining, fine-tuning, dan alignment), ekstensi multimodal, teknik engineering prompt, penalaran berantai (chain-of-thought), agen, serta topik terkini seperti keamanan model dan perlindungan privasi.
Gambaran Umum Kursus
📚 Ringkasan Konten
Kursus ini menyediakan pengantar komprehensif dan mendalam mengenai evolusi Model Bahasa Besar (LLM), arsitektur teknis inti, paradigma pelatihan (pre-training, fine-tuning, dan alignment), ekstensi multimodal, teknik prompt engineering, Chain of Thought (CoT), agen, serta topik terdepan seperti keamanan model dan perlindungan privasi.
Analisis mendalam mengenai evolusi teknologi dan keselarasan keamanan model besar secara menyeluruh, mulai dari pre-training hingga agen umum.
🎯 Tujuan Pembelajaran
- Membedakan antara arsitektur model: Mengidentifikasi perbedaan struktural dan kasus penggunaan untuk model Encoder-only, Decoder-only, dan Encoder-Decoder.
- Menjelaskan Pipeline Pelatihan LLM: Mendeskripsikan transisi dari pre-training berbasis self-supervised ke Supervised Fine-Tuning (SFT) dan Reinforcement Learning from Human Feedback (RLHF).
- Menganalisis Skala dan Perilaku Model: Menjelaskan konsep Hukum Skala, Kemampuan Muncul (pembelajaran dalam konteks, Chain of Thought), serta fenomena Halusinasi.
- Menganalisis perbedaan struktur antara arsitektur Encoder-only (BERT), Decoder-only (GPT), dan Encoder-Decoder (T5).
- Menjelaskan proses pelatihan tiga tahap: Pre-training (model dasar), Instruction Tuning (SFT), dan Alignment (RLHF/PPO).
- Membandingkan kinerja, hukum skala, dan inovasi arsitektur dari LLM utama seperti GPT, Llama, Qwen, dan DeepSeek.
- Menerapkan strategi prompting zero-shot dan few-shot untuk ekstraksi dan klasifikasi data terstruktur.
- Mengkalibrasi hipertes (Temperature, Top P, Penalti) untuk menyeimbangkan output kreatif dan deterministik.
- Membangun prompt Chain-of-Thought (CoT) yang efektif menggunakan metode manual, otomatis, dan zero-shot ("Mari kita pikir langkah demi langkah").
- Menganalisis dan Membandingkan Varian CoT: Membedakan antara Self-Consistency, Program of Thought (PoT), Tree-of-Thought (ToT), dan Graph-of-Thought (GoT).
Pelajaran 共 8 课时 · 预计 24.0h
Pelajaran
Lesson
This lesson explores the paradigm shift from task-specific AI to general-purpose Large Language Models, highlighting how scaling laws and the Transformer architecture enable emergent reasoning capabilities. Students will learn how the self-attention mechanism allows models to process data in parallel and why decoder-only architectures have become the standard for modern generative AI.
This lesson explores the evolution of Transformer architectures, highlighting why the industry has shifted toward Decoder-only models due to their superior scaling laws and generative capabilities. It also examines the foundational training pipeline, covering the transition from pre-training and Chinchilla optimality to modern instruction tuning and alignment techniques.
This lesson explores the evolution of prompting strategies, ranging from zero-shot instructions to few-shot demonstrations and Chain-of-Thought reasoning for complex logic. Students will learn how to improve model performance and reliability by using structural constraints, logical decomposition, and precise formatting to ensure outputs are suitable for programmatic use.
This lesson explores the evolution of reasoning in large language models, moving from linear Chain-of-Thought to advanced structured architectures like Tree-of-Thought and Graph-of-Thought. Students will learn how these frameworks, alongside techniques like Program of Thought and knowledge editing, enable models to perform deliberate planning, verification, and complex mathematical problem-solving.
This lesson explores the challenges of AI safety, focusing on the training paradox between model helpfulness and harmlessness, as well as the mechanics of adversarial jailbreak attacks. It also introduces statistical watermarking techniques, such as the KGW framework, which use vocabulary partitioning and logit bias to embed detectable, invisible signatures into AI-generated content.
This lesson explores the fundamental conflict in AI alignment between maintaining model helpfulness and ensuring harmlessness, highlighting how jailbreak attacks exploit this tension through role-playing and cognitive coercion. Students will learn how these vulnerabilities function and examine technical strategies, such as intent analysis and robust system prompting, to defend against prompt-based bypass attempts.
This lesson explores the evolution of Multi-modal Large Language Models (MLLMs) from vision-centric systems to integrated architectures capable of processing diverse sensory inputs like audio and 3D data. Students will learn how specialized encoders and projection bridges align non-textual signals into a unified semantic space, enabling the model to effectively reason across multiple modalities.
This lesson explores the development of autonomous GUI agents that use a tripartite architecture—planning, decision-making, and reflection—to interact with software interfaces. It further examines how Reinforcement Learning and RLHF enable these agents to adapt to dynamic environments while maintaining safety and reliability against potential risks.