AI Dictionary › AI Fundamentals

Greedy decoding

Decodifica greedy

Greedy decoding is the simplest strategy for generating text: at each step the model always picks the highest-probability token, with no hesitation or alternatives. It is like a hiker who at every fork always takes the path that looks most downhill right now, never weighing the bigger picture. The result is deterministic: the same request always yields the same identical answer. This makes it fast and predictable but short-sighted: a locally optimal choice can lead to a worse overall sentence, because it never explores paths that start badly and end better.

Definition

Greedy decoding matters as the most basic baseline, used when reproducible answers are needed, as in tests or factual tasks. The downside is a tendency toward flat, repetitive, uncreative text.

Related terms

• Sampling
• Beam search
• Top-p (nucleus sampling)
• Temperature
• Top-k

More in AI Fundamentals

• LLM (Large Language Model)
• Token
• Context Window
• Temperature
• AI Hallucination
• Fine-tuning
• RAG (Retrieval Augmented Generation)
• Embedding

Put it into practice

• Real scenario: apply this concept with an instant AI verdict.
• Daily gym: short drills to train dictionary concepts.
• Playoff game: challenge other players, elimination-style.

📱 Download the Android app (beta) iOS coming soon

Say what you mean. Get what you need.

Grace Certified, the AI coach that trains and certifies your prompt engineering, by Agora Intelligence.