AIQASolver
Glossary

The AI & LLM Glossary

Practitioner-grade definitions for AI evaluation, RAG, safety, and monitoring — written for engineers shipping real systems.

Evaluation

LLM Evaluation

The process of systematically measuring the quality, safety, and reliability of large language model outputs.

LLM-as-a-Judge

Using a capable LLM to score outputs of another LLM against a rubric.

RAG

RAG Evaluation

Measuring retrieval-augmented generation systems across both the retrieval and generation stages.

Safety

Hallucination

An LLM output that is fluent and confident but factually wrong or unsupported by the provided context.

AI Red Teaming

Adversarial testing of AI systems to surface safety, security, and reliability failures before users do.

Prompt Injection

An attack where untrusted input overrides or manipulates the model's original instructions.

Metrics

BLEU Score

An n-gram precision metric originally built for machine translation, comparing model output to one or more reference texts.

ROUGE Score

A recall-oriented family of metrics that measures overlap between generated text and reference summaries.

Perplexity

A measure of how 'surprised' a language model is by a sequence — lower is better.

Monitoring

Model Drift

Degradation in model performance over time as production data diverges from training or evaluation data.