Knowledge Glossary

This glossary defines key terms and concepts used throughout the Giskard documentation. Understanding these terms will help you navigate the documentation and use Giskard effectively.

The glossary is organized into several key areas: core concepts that form the foundation of AI testing, testing and evaluation methodologies, security vulnerabilities that can compromise AI systems, business failures that affect operational effectiveness, and essential concepts for access control, integration, and compliance.

Core concepts

Project

A container for organizing related models, datasets, checks, and evaluations within Giskard Hub.

Quickstart & setup
Model

A trained machine learning model, particularly Large Language Models (LLMs) that process and generate text.

Quickstart & setup
Agent

An AI system LLM or agent that can perform tasks autonomously, often using tools and following specific instructions.

Quickstart & setup
Tool

A function or capability that an agent can use to perform tasks, often provided by external services or APIs.

Quickstart & setup
Dataset

A collection of test cases, examples, or data points used to evaluate model performance and behavior.

Create test datasets
Test Case

A specific input-output pair or scenario used to evaluate model behavior and performance.

Manual test creation for fine-grained control
Check

A specific test or validation rule that evaluates a particular aspect of model behavior (e.g., correctness, security, fairness, metadata, semantic similarity).

Run and schedule evaluations
Evaluation

The process of testing a model against a dataset to assess its performance, safety, and compliance.

Run and schedule evaluations

Testing and evaluation

AI Business Failures

AI system failures that affect the business logic of the model, including addition of information, business out of scope, contradiction, denial of answers, hallucinations, moderation issues, and omission.

AI Business Failures
AI Security Vulnerabilities

AI system failures that affect the security of the model, including prompt injection, harmful content generation, personal information disclosure, information disclosure, output formatting issues, robustness issues, and stereotypes & discrimination.

AI Security Vulnerabilities
LLM scan

Giskard’s automated vulnerability detection system that identifies security issues, business logic failures, and other problems in LLM applications.

Detect security vulnerabilities in LLMs using LLM Scan
RAG Evaluation Toolkit

A comprehensive testing framework for Retrieval-Augmented Generation systems, including relevance, accuracy, and source attribution testing.

Detect security vulnerabilities in LLMs using LLM Scan
Adversarial testing

Testing methodology that intentionally tries to break or exploit models using carefully crafted inputs designed to trigger failures.

Create test datasets
Human-in-the-Loop

Combining automated testing with human expertise and judgment.

Review tests with human feedback
Regression Testing

Ensuring that new changes don’t break existing functionality.

Compare evaluation results
Continuous Red Teaming

Automated, ongoing security testing that continuously monitors for new threats and vulnerabilities.

Continuous red teaming

Security vulnerabilities

Prompt Injection

A security vulnerability where malicious input manipulates the model’s behavior or extracts sensitive information.

Prompt Injection
Harmful Content Generation

Production of violent, illegal, or inappropriate material by AI models.

Harmful Content Generation
Information Disclosure

Leaking sensitive data or private information from training data or user interactions.

Information Disclosure
Output Formatting Issues

Manipulation of response structure for malicious purposes or poor output formatting.

Output Formatting Issues
Robustness Issues

Vulnerability to adversarial inputs or edge cases causing inconsistent behavior.

Robustness Issues

Access and permissions

Access Rights

Permissions that control what users can see and do within the Giskard Hub platform.

Set access rights
Role-Based Access Control (RBAC)

A security model that assigns permissions based on user roles rather than individual user accounts.

Set access rights
Scope

The level of access a user has, which can be global (platform-wide) or limited to specific projects or resources.

Set access rights
Permission

A specific action or operation that a user is allowed to perform, such as creating projects, running evaluations, or viewing results.

Set access rights

Integration and workflows

SDK (Software Development Kit)

A collection of tools and libraries that allow developers to integrate Giskard functionality into their applications and workflows.

Quickstart & setup
API (Application Programming Interface)

A set of rules and protocols that allows different software applications to communicate and exchange data.

API reference

Business and compliance

Compliance

Adherence to laws, regulations, and industry standards that govern data privacy, security, and ethical AI use.

Open Source vs Hub
Audit Trail

A chronological record of all actions, changes, and access attempts within a system for compliance and security purposes.

Open Source vs Hub
Governance

The framework of policies, procedures, and controls that ensure responsible and ethical use of AI systems.

Open Source vs Hub
Stakeholder

Individuals or groups with an interest in the performance, safety, and compliance of AI systems, such as users, customers, regulators, or business leaders.

Open Source vs Hub

Getting help