APPENDIX F: FURTHER READING AND RESOURCES

ARTICLES AND PAPERS Resource: Building effective agents, an essay published by Anthropic in December 2024. Where to find it: anthropic.com/research/building-effective-agents Why read it: it explains the difference between workflows and agents and advises starting with the simplest solution, which matches the advice in Appendix C. Resource: ReAct: Synergizing Reasoning and Acting in Language Models, by Shunyu Yao and colleagues. Where to find it: arxiv.org/abs/2210.03629 Why read it: it is the research paper behind the ReAct pattern taught in Chapter 9. Resource: Judging LLM-as-a-Judge with Multi-Turn Benchmark (MT-Bench) and Chatbot Arena, by Lianmin Zheng and colleagues. Where to find it: arxiv.org/abs/2306.05685 Why read it: it studies using a model as a grader and discusses its biases, which supports Chapter 33 and Appendix E. Resource: Lost in the Middle: How Language Models Use Long Contexts, by Nelson F. Liu and colleagues. Where to find it: arxiv.org/abs/2307.03172 Why read it: it reports that models use information at the start and end of a long context better than information in the middle, which supports Chapters 19 and 20. Resource: AgentOps: Enabling Observability of LLM Agents, by Liming Dong, Qinghua Lu, and Liming Zhu. Where to find it: arxiv.org/abs/2411.05285 Why read it: it proposes a taxonomy of what to record when watching agents, which supports Chapters 27 through 29 and 59. Resource: Survey on Evaluation of LLM-based Agents, led by Asaf Yehudai. Where to find it: arxiv.org/abs/2503.16416 Why read it: it surveys how agents are tested and points out gaps such as measuring cost, safety, and robustness, supporting Chapters 31 through 34.