Fonti

Fonti Autorevoli

Le definizioni, i pattern e le pratiche di questo pack sono basati sulle seguenti fonti primarie e ampiamente riconosciute. Dove un concetto ha un originatore canonico (per esempio ReAct o il Model Context Protocol), si cita l'articolo originale o la documentazione mantenuta.

Guida su Agenti e Tool-Use

  • Building Effective Agents - Anthropic Engineering. https://www.anthropic.com/research/building-effective-agents
  • Tool use (function calling) - documentazione API Anthropic. https://docs.anthropic.com/en/docs/build-with-claude/tool-use
  • Function calling - documentazione OpenAI Platform. https://platform.openai.com/docs/guides/function-calling
  • Agents - documentazione OpenAI Platform. https://platform.openai.com/docs/guides/agents

Protocolli e Standard

  • Model Context Protocol - specifica e documentazione ufficiale. https://modelcontextprotocol.io/
  • Model Context Protocol - Introduzione (Anthropic). https://www.anthropic.com/news/model-context-protocol

Ricerca Fondazionale

  • ReAct: Synergizing Reasoning and Acting in Language Models - Yao et al., 2022. https://arxiv.org/abs/2210.03629
  • Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks - Lewis et al., 2020. https://arxiv.org/abs/2005.11401
  • Chain-of-Thought Prompting Elicits Reasoning in Large Language Models - Wei et al., 2022. https://arxiv.org/abs/2201.11903
  • Tree of Thoughts: Deliberate Problem Solving with Large Language Models - Yao et al., 2023. https://arxiv.org/abs/2305.10601
  • Efficient and robust approximate nearest neighbor search using HNSW graphs - Malkov & Yashunin, 2016. https://arxiv.org/abs/1603.09320
  • LoRA: Low-Rank Adaptation of Large Language Models - Hu et al., 2021. https://arxiv.org/abs/2106.09685
  • Direct Preference Optimization - Rafailov et al., 2023. https://arxiv.org/abs/2305.18290

Prompting e Pratiche di Ingegneria

  • Prompt engineering overview - documentazione Anthropic. https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering/overview
  • Prompt caching - documentazione Anthropic. https://docs.anthropic.com/en/docs/build-with-claude/prompt-caching
  • OpenAI Cookbook - ricette pratiche per RAG, agenti ed eval. https://cookbook.openai.com/

Framework e Strumenti

  • Documentazione LangGraph (orchestrazione di agenti). https://langchain-ai.github.io/langgraph/
  • Documentazione LlamaIndex (RAG e agenti). https://docs.llamaindex.ai/
  • OpenAI Agents SDK. https://openai.github.io/openai-agents-python/
  • pgvector - ricerca di similarità vettoriale open-source per PostgreSQL. https://github.com/pgvector/pgvector

Valutazione

  • RAGAS - framework di valutazione per pipeline RAG. https://docs.ragas.io/
  • promptfoo - test e valutazione di app LLM. https://www.promptfoo.dev/docs/intro/
  • DeepEval - framework di valutazione LLM. https://docs.confident-ai.com/

Sicurezza

  • OWASP Top 10 for Large Language Model Applications. https://owasp.org/www-project-top-10-for-large-language-model-applications/