Generative AI & LLMs
Emergent capability
A useful behavior that appears as models scale or train, even when it was not explicitly programmed.
Example
A large model unexpectedly performs simple translation even though translation was not its explicit training task.
Why it matters
It gives researchers a way to discuss behavior they did not specifically teach or expect.
You'll hear it
“It handled that task even though we hadn't planned for it.”
Can you control it?
Not directly
No, not directly. Emergent capability describes a learned property, observed result, or broader concept rather than a normal user setting. You may influence it indirectly by changing the tool, data, prompt, or workflow.
Common questions
- Does one surprising answer establish a new capability?
- No. Accidental success or a familiar example can look more impressive than the broader ability deserves.
- Can scoring affect when an ability seems to appear?
- Yes. A strict pass-or-fail score can make gradual improvement look sudden.
- Does an unexpected ability imply consciousness?
- No. A newly observed skill does not establish awareness or personal experience.