Existing methods for adapting large language models (LLMs) to new tasks are
not suited to multi-task adaptation because they modify all the model weights
-- causing destructive interference between tasks. The resulting effects, such
as catastrophic forgetting of earlier tasks, make it challenging to obtain good
performance on multiple tasks at the same time. To mitigate this, we propose
Lottery Ticket Adaptation (LoTA), a sparse adaptation method that identifies
and optimizes only a sparse subnetwork of the model. We evaluate LoTA on a wide
range of challenging tasks such as instruction following, reasoning, math, and
summarization. LoTA obtains better performance than full fine-tuning and
low-rank adaptation (LoRA), and maintains good performance even after training
on other tasks -- thus, avoiding catastrophic forgetting. By extracting and
fine-tuning over \emph{lottery tickets} (or \emph{sparse task vectors}), LoTA
also enables model merging over highly dissimilar tasks.

Lottery Ticket Adaptation (LoTA) is a sparse adaptation method for large language models that improves performance on multiple tasks by optimizing a sparse subnetwork, mitigating catastrophic forgetting.