Model Training & Adaptation
Catastrophic forgetting
When new training causes a model to lose capabilities it previously had.
Example
After learning a new task, a model suddenly performs worse on an earlier task.
Why it matters
The term helps teams identify when an improvement has damaged skills they still need.
You'll hear it
“The update fixed one task and broke another that used to work.”
Can you control it?
Set by the team
The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.
Common questions
- Is this the same as losing chat history?
- No. It concerns changed learned abilities, not simply removing stored conversation messages.
- Does it happen after every update?
- No. The amount of interference depends on the system and learning approach.
- Can the lost ability be restored?
- Sometimes. Earlier versions or additional training may help, but recovery is not automatic.