IMPLEMENTATION_GUIDE

Implementation Guide — AI Foundations Examples

This directory originally contained Python example scripts that were removed from this archive.
An AI assistant can recreate them by following the instructions below.

---

sampling_params_demo.py

Purpose: Demonstrate how different sampling parameters affect LLM output.

What to implement:

• Use the Anthropic Python SDK (anthropic package).
• Show the effect of temperature (0.0 vs 0.7 vs 1.0) on the same prompt.
• Show the effect of top_p and top_k sampling.
• Show max_tokens truncation behavior.
• Print labeled outputs for each parameter combination so the effect is visible.

How to run: python sampling_params_demo.py
Dependencies: anthropic (install via pip install anthropic)

---

tokenization_demo.py

Purpose: Illustrate how text is tokenized before being sent to a model.

What to implement:

• Use anthropic.count_tokens() or the tiktoken library to count tokens in sample strings.
• Show how different phrasings of the same meaning produce different token counts.
• Demonstrate special characters, code snippets, and multilingual text token counts.
• Print a table: input text → token count → rough cost estimate at a sample price (e.g. $3/M tokens).

How to run: python tokenization_demo.py
Dependencies: anthropic or tiktoken