Leaking LoRa: An Evaluation of Password Leaks and Knowledge Storage in Large Language Models
Fuente:
arXiv
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908292381409280 |
|---|---|
| author | Marinelli, Ryan Eckhoff, Magnus |
| author_facet | Marinelli, Ryan Eckhoff, Magnus |
| contents | To effectively deploy Large Language Models (LLMs) in application-specific settings, fine-tuning techniques are applied to enhance performance on specialized tasks. This process often involves fine-tuning on user data data, which may contain sensitive information. Although not recommended, it is not uncommon for users to send passwords in messages, and fine-tuning models on this could result in passwords being leaked. In this study, a Large Language Model is fine-tuned with customer support data and passwords from the RockYou password wordlist using Low-Rank Adaptation (LoRA). Out of the first 200 passwords from the list, 37 were successfully recovered. Further, causal tracing is used to identify that password information is largely located in a few layers. Lastly, Rank One Model Editing (ROME) is used to remove the password information from the model, resulting in the number of passwords recovered going from 37 to 0. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_00031 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Leaking LoRa: An Evaluation of Password Leaks and Knowledge Storage in Large Language Models Marinelli, Ryan Eckhoff, Magnus Cryptography and Security Artificial Intelligence Computation and Language To effectively deploy Large Language Models (LLMs) in application-specific settings, fine-tuning techniques are applied to enhance performance on specialized tasks. This process often involves fine-tuning on user data data, which may contain sensitive information. Although not recommended, it is not uncommon for users to send passwords in messages, and fine-tuning models on this could result in passwords being leaked. In this study, a Large Language Model is fine-tuned with customer support data and passwords from the RockYou password wordlist using Low-Rank Adaptation (LoRA). Out of the first 200 passwords from the list, 37 were successfully recovered. Further, causal tracing is used to identify that password information is largely located in a few layers. Lastly, Rank One Model Editing (ROME) is used to remove the password information from the model, resulting in the number of passwords recovered going from 37 to 0. |
| title | Leaking LoRa: An Evaluation of Password Leaks and Knowledge Storage in Large Language Models |
| topic | Cryptography and Security Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2504.00031 |