LM-Fix: Lightweight Bit-Flip Detection and Rapid Recovery Framework for Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Tahmasivand, Ahmad, Zahran, Noureldin, Al-Sayouri, Saba, Fouda, Mohammed, Khasawneh, Khaled N.
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912922385514496
author Tahmasivand, Ahmad
Zahran, Noureldin
Al-Sayouri, Saba
Fouda, Mohammed
Khasawneh, Khaled N.
author_facet Tahmasivand, Ahmad
Zahran, Noureldin
Al-Sayouri, Saba
Fouda, Mohammed
Khasawneh, Khaled N.
contents This paper presents LM-Fix, a lightweight detection and rapid recovery framework for faults in large language models (LLMs). Existing integrity approaches are often heavy or slow for modern LLMs. LM-Fix runs a short test-vector pass and uses hash-guided checks to detect bit-flip faults, then repairs them locally without a full reload. Across multiple models, it detects over 94% of single-bit flips at TVL=200 and nearly 100% of multi-bit flips with approximately 1% to 7.7% runtime overhead; recovery is more than 100x faster than reloading. These results show a practical, low-overhead solution to keep LLMs reliable in production
format Preprint
id arxiv_https___arxiv_org_abs_2511_02866
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle LM-Fix: Lightweight Bit-Flip Detection and Rapid Recovery Framework for Language Models
Tahmasivand, Ahmad
Zahran, Noureldin
Al-Sayouri, Saba
Fouda, Mohammed
Khasawneh, Khaled N.
Software Engineering
Artificial Intelligence
Hardware Architecture
Cryptography and Security
This paper presents LM-Fix, a lightweight detection and rapid recovery framework for faults in large language models (LLMs). Existing integrity approaches are often heavy or slow for modern LLMs. LM-Fix runs a short test-vector pass and uses hash-guided checks to detect bit-flip faults, then repairs them locally without a full reload. Across multiple models, it detects over 94% of single-bit flips at TVL=200 and nearly 100% of multi-bit flips with approximately 1% to 7.7% runtime overhead; recovery is more than 100x faster than reloading. These results show a practical, low-overhead solution to keep LLMs reliable in production
title LM-Fix: Lightweight Bit-Flip Detection and Rapid Recovery Framework for Language Models
topic Software Engineering
Artificial Intelligence
Hardware Architecture
Cryptography and Security
url https://arxiv.org/abs/2511.02866