FT2: First-Token-Inspired Online Fault Tolerance on Critical Layers for Generative Large Language Models
Open Access
- 20 July 2025
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 1-14
Abstract
No abstract availableFunding Information
- NSF (National Science Foundation) (2410856, 2402940)
- DOE U.S. Department of Energy (66150, DE-AC05-76RL01830)
- CCI (Commonwealth Cyber Initiative) (HC?3Q24?047)
This publication has 27 references indexed in Scilit:
- Understanding Mixed Precision GEMM with MPGemmFI: Insights into Fault Resilience2024
- A Low-cost Fault Corrector for Deep Neural Networks through Range Restriction2021
- Enabling Software Resilience in GPGPU Applications via Partial Thread Protection2021
- TensorFI: A Flexible Fault Injection Framework for TensorFlow Applications2020
- PyTorchFI: A Runtime Perturbation Tool for DNNs2020
- Practical Resilience Analysis of GPGPU Applications in the Presence of Single- and Multi-Bit FaultsIEEE Transactions on Computers, 2020
- Bit-Flip Attack: Crushing Neural Network With Progressive Bit Search2019
- Code-Dependent and Architecture-Dependent Reliability Behaviors2018
- ImageNet: A large-scale hierarchical image database2009
- Soft Errors in Advanced Computer SystemsIEEE Design & Test of Computers, 2005