Resources

Course Resources

This page collects the tools, frameworks, benchmarks, and foundational papers referenced throughout the course textbook, Agentic Large Language Models, organized by course part.

---

The Textbook

• Agentic Large Language Models — Max van Duijn, Michiel van der Meer, Aske Plaat, Niki van Stein. The primary reading material for this course; see the Syllabus for the chapter-by-chapter breakdown.

---

Development Environment

• Python — primary programming language for all assignments
• PyTorch — deep learning framework used throughout the case studies
• Hugging Face — models, datasets, and libraries (transformers, model cards, hub) used in Case Study 1 and beyond
• Jupyter Notebooks / Google Colab — interactive development

---

Part I — Language Modeling

• Vaswani et al. (2017) — Attention Is All You Need
• nanochat — the from-scratch ChatGPT-style training pipeline used in Case Study 1

Part II — Reasoning

• Wei et al. (2022) — Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
• Yao et al. (2022) — ReAct: Synergizing Reasoning and Acting in Language Models
• Shao et al. (2024) — DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models (GRPO)
• Guo et al. (2025) — DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
• Kim et al. (2024) — OpenVLA: An Open-Source Vision-Language-Action Model
• Black et al. (2024) — π₀: A Vision-Language-Action Flow Model for General Robot Control

Part III — Agents

• BALROG — benchmark suite (NetHack, TextWorld, and other games/puzzles) used in Case Study 2 to evaluate reasoning and reflective agents
• ReAct / Reflexion — reference implementations for the self-reflective agent loop

Part IV — Tools for Agentic Systems

• Model Context Protocol (MCP) — standardized tool-integration protocol discussed in the tools chapter and used in Case Study 3
• LangChain — framework for toolchains and the ReAct loop
• Lewis et al. (2020) — Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
• Self-RAG and Self-Ask — adaptive retrieval methods covered in the retrieval chapter

Part V — Behavioral and Cognitive Perspectives

• Park et al. (2023) — Generative Agents: Interactive Simulacra of Human Behavior — framework used in Case Study 4
• CAMEL / AutoGen / CrewAI — multi-agent orchestration frameworks discussed in the role-based interaction chapter

Part VI — Advanced Topics

• Nanda et al. (2023) — Progress Measures for Grokking via Mechanistic Interpretability

---

Getting Help

• Office Hours: Check the Schedule for lecture times; office hours are announced in class
• Email: a.plaat@liacs.leidenuniv.nl for course-related questions
• LIACS: Visit the LIACS website for general support

---

Last updated: July 2026