Safety, Guardrails & Responsible AI

89 lessons in AI Engineering / Working with LLMs

• Guardrail FrameworksQuiz
• Copyright and IP RisksQuiz
• Transparency and AI DisclosureQuiz
• Red-Teaming LLM ApplicationsQuiz
• Logging and AuditabilityQuiz
• Rate Limiting and Abuse PreventionQuiz
• Grounding to Reduce Unsafe OutputsQuiz
• Sensitive-Topic RestrictionsQuiz
• Data Privacy and ComplianceQuiz
• Responsible AI PrinciplesQuiz
• RLHF and Alignment BasicsQuiz
• Output Validation and FilteringQuiz
• Human-in-the-Loop ReviewQuiz
• Refusal and Over-RefusalQuiz
• System Prompt HardeningQuiz
• Hallucination as a Safety RiskQuiz
• Toxicity and Bias in OutputsQuiz
• PII Detection and RedactionQuiz
• Categories of Harmful ContentQuiz
• Content Moderation APIsQuiz
• Jailbreaking and DefensesQuiz
• Prompt Injection AttacksQuiz
• Input vs Output GuardrailsQuiz
• What Guardrails AreQuiz
• Human-in-the-Loop Review and EscalationSlides / Video
• Guardrail Frameworks and LibrariesSlides / Video
• Bias and Fairness in LLM OutputsSlides / Video
• Handling Hallucinations ResponsiblySlides / Video
• PII Detection and RedactionSlides / Video
• Topic and Scope RestrictionSlides / Video
• Blocklists, Allowlists and Keyword FilteringSlides / Video
• System Prompt Hardening for SafetySlides / Video
• Moderation APIs and Content ClassifiersSlides / Video
• Output Filtering and ModerationSlides / Video
• Input Validation for User PromptsSlides / Video
• Handling Model RefusalsCode Quiz
• Jailbreaking and How Models Are BypassedSlides / Video
• Prompt Injection AttacksSlides / Video
• Categories of LLM Safety RisksSlides / Video
• Prompt Injection DetectionCode Quiz
• Safe Fallback on GuardrailCode Quiz
• Topic Allow-List ScopeCode Quiz
• Redacting PIICode Quiz
• Logging Flagged RequestsCode Quiz
• Reading Flagged CategoriesCode Quiz
• Two-Pass Moderation FlowCode Quiz
• Preventing System Prompt LeakageCode Quiz
• Grounding to Source ContextCode Quiz
• Validating Output SchemaCode Quiz
• Calling the Moderation APICode Quiz
• LLM-as-Judge GuardrailCode Quiz
• Denylist Keyword FilterCode Quiz
• Sanitizing Untrusted InputCode Quiz
• System Safety Instruction RoleCode Quiz
• max_tokens Output LimitCode Quiz
• Rate Limiting RequestsCode Quiz
• Blocking Flagged ResponsesCode Quiz
• Delimiting Untrusted ContentCode Quiz
• What Guardrails Are and Why They MatterSlides / Video
• Data Privacy in LLM AppsSlides / Video
• Logging and Monitoring for AI SafetySlides / Video
• Rate Limiting and Abuse PreventionFlashcard
• Output Filtering and ModerationFlashcard
• Prompt Injection AttacksFlashcard
• PII Detection and RedactionFlashcard
• Refusals and Safe Completion BehaviorFlashcard
• Topic and Scope RestrictionFlashcard
• Hallucination Mitigation StrategiesFlashcard
• Human-in-the-Loop OversightFlashcard
• Model Cards and TransparencyFlashcard
• Toxicity and Harmful Content DetectionFlashcard
• Guardrail Frameworks and LibrariesFlashcard
• Red Teaming LLM ApplicationsFlashcard
• Audit Logging and TraceabilityFlashcard
• What Guardrails Are and Why They MatterFlashcard
• Input Validation and Prompt FilteringFlashcard
• Content Moderation APIsFlashcard
• Jailbreaking and Adversarial PromptsFlashcard
• Defending Against Prompt InjectionFlashcard
• Data Privacy and ComplianceFlashcard
• Bias and Fairness in LLM OutputsFlashcard
• Grounding Responses to Reduce FabricationFlashcard
• System Prompt HardeningFlashcard
• Structured Output Constraints for SafetyFlashcard
• Responsible AI PrinciplesFlashcard
• Data Leakage and Training Data ExposureFlashcard
• Responsible AI PrinciplesSlides / Video
• Layered Defense: Combining Multiple GuardrailsSlides / Video
• Refusal Behavior and Graceful DeclinesSlides / Video