Practice: Prompt Engineering Mastery

# Practice: Evaluate the quality of AI-generated text
# Identify hallucinations, inconsistencies, and factual issues

ai_generated_samples = [
    {
        "prompt": "What year was Python programming language created?",
        "generated": "Python was created by Guido van Rossum and first released in 1991.",
        "correct": True,
        "note": "Accurate. Python was indeed released in 1991.",
    },
    {
        "prompt": "Who wrote the Transformer paper?",
        "generated": "The Transformer was invented by Yann LeCun at NYU in 2015.",
        "correct": False,
        "note": "Hallucination. The Transformer was introduced by Vaswani et al. at Google in 2017.",
    },
    {
        "prompt": "What is the capital of Australia?",
        "generated": "The capital of Australia is Canberra.",
        "correct": True,
        "note": "Accurate. Canberra is the capital, though Sydney is larger and more famous.",
    },
    {
        "prompt": "How many GPT-3 parameters?",
        "generated": "GPT-3 has 175 trillion parameters, making it the largest language model ever built.",
        "correct": False,
        "note": "Hallucination. GPT-3 has 175 billion parameters, not trillion. Off by 1000x.",
    },
]

print("Evaluating AI-Generated Text for Accuracy:")
print()

correct = 0
for sample in ai_generated_samples:
    status = "CORRECT" if sample["correct"] else "HALLUCINATION"
    print(f"Prompt: '{sample['prompt']}'")
    print(f"  Generated: '{sample['generated']}'")
    print(f"  Status: {status}")
    print(f"  Note: {sample['note']}")
    print()
    if sample["correct"]: correct += 1

accuracy = correct / len(ai_generated_samples) * 100
print(f"Accuracy in this sample: {accuracy:.0f}% ({correct}/{len(ai_generated_samples)} correct)")
print()
print("This illustrates why verifying AI-generated facts is important.")