# Catastrophic Forgetting

> A model learns a new task and loses skills it had.

**Field:** Artificial Intelligence  
**Category:** LLMs  
**Also known as:** Catastrophic Interference, Forgetting After Fine-Tuning

## Think of it like

It is like you learn a new song so well that the old one is gone.

## Definition

Catastrophic forgetting is when a model learns a new task and gets worse at old ones. Training changes the same weights that held the old skills, so new learning can write over them. In language models it often shows up after tuning on a narrow set of data. A model tuned only on legal text may get worse at math or chat. Mixing some old data back in, or changing only a few weights, can lower the risk.

## Example

Jack fine-tunes a chat model on his shop's help tickets. It now answers tickets well, but it starts to fail simple sums it used to get right.

## Why it matters

A tuned model can pass its new test and still fail the old ones. You have to check both.

## Common mistake

Testing only the new task. Run the old tests too, or you will not see what was lost.

## Three principles

- **Same weights.** New learning writes over old.
- **Narrow data.** Raises the risk.
- **Mix and limit.** Old data and fewer changed weights help.

## Related terms

- [Fine-Tuning](https://learn.benibix.com/term/ai/fine-tuning)
- [Parameter-Efficient Fine-Tuning (PEFT)](https://learn.benibix.com/term/ai/parameter-efficient-fine-tuning)
- [Low-Rank Adaptation (LoRA)](https://learn.benibix.com/term/ai/low-rank-adaptation)
- [Instruction Tuning](https://learn.benibix.com/term/ai/instruction-tuning)
- [Model Collapse](https://learn.benibix.com/term/ai/model-collapse)

## Told apart from

- [Catastrophic Forgetting vs Model Collapse](https://learn.benibix.com/compare/catastrophic-forgetting-vs-model-collapse): People think catastrophic forgetting and model collapse are one problem, because both make a model worse after training. They are not. Forgetting happens inside one model: new training writes over skills it had. Collapse builds up across many models, each one trained on text made by the one before. The fix for the first is to test old skills. The fix for the second is to keep real human data.

## Taught in

- [Fine-Tuning and LoRA](https://learn.benibix.com/learn/ai-llms/fine-tuning-and-lora): LLMs, 9/15

![Two sets of skill bars for one model. Before fine-tuning, chat, math and legal text are all medium. After fine-tuning on legal text, the legal bar is tall, while chat and math have dropped.](https://learn.benibix.com/diagrams/ai9-catastrophic-forgetting.svg)

---

Source: https://learn.benibix.com/term/ai/catastrophic-forgetting
Author: Behnia Zarei (a developer in Vienna, Austria, who builds and teaches data, AI and business intelligence at BENIBIX)
Publisher: BENIBIX Learn (https://learn.benibix.com/)
Last updated: 2026-11-18
Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote freely with attribution.
