1 / 8
LLMs in practice: capabilities, limits, hallucinations
In this lesson
LLMs in practice: capabilities, limits, hallucinations
Identify the advantages and disadvantages of generative AI solutions in production contexts, including hallucinations, nondeterminism, and interpretability challenges [^3].
You'll be able to
- Identify the advantages and disadvantages of generative AI solutions in production contexts, including hallucinations, nondeterminism, and interpretability challenges [^3].
- Classify grounding techniques (retrieval-augmented generation, first-party enterprise data, third-party data, and world data) and explain how they affect LLM-generated output to mitigate misinformation [^1].
- Apply sampling parameters and settings (token count, temperature, top-p nucleus sampling, safety settings, and output length) to control the behavior of generative AI models in real-world deployments [^1].
- Evaluate LLM vulnerabilities from the OWASP Top 10 for LLM Applications (2025), including prompt injection, misinformation, and improper output handling, and propose appropriate mitigations [^2][^4].
- Assess fine-tuning impact and select parameter-efficient update techniques to reduce hallucinations and improve model reliability in NVIDIA-aligned workflows [^5].