AI Glossary
5324+ terms covering machine learning, neural networks, LLMs, prompting, frameworks, and AI safety. From fundamentals to advanced concepts.
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
Z
Zephyr
A series of fine-tuned open LLMs from Hugging Face optimized for helpful assistants. Demonstrates how smaller models with good alignment can outperform larger base models.
ZeRO Optimization
Zero Redundancy Optimizer, a distributed training technique that partitions optimizer states, gradients, and parameters across data-parallel processes to reduce memory redundancy.
Zero-Day (AI Context)
Novel vulnerabilities or attack vectors discovered in AI systems before developers know about them. AI security research increasingly focuses on finding such vulnerabilities proactively.
Learn about AI safety →Zero-Shot Chain-of-Thought
Adding "Let's think step by step" to a prompt to trigger reasoning without providing examples. A simple but effective technique for improving accuracy on complex tasks.
Explore all frameworks →Zero-Shot Learning
When AI performs a task without examples in the prompt, relying entirely on pre-trained knowledge and instructions. Contrasts with few-shot where examples are provided.
Explore all frameworks →Zero-Shot NER
Named entity recognition performed on entity types not seen during training, using natural language descriptions of entity categories to generalize to new entity types without additional labeled data.
Zero-Shot Object Detection
The ability to detect and localize objects of novel categories without any training examples, using vision-language alignment to match text descriptions of new categories to visual regions.
Put These Terms Into Practice
Now that you know the vocabulary, learn the methodologies that make AI work for you.