Text generated by OPUS-MT and T5 models with single-bit errors in the parameters

Description

The dataset contains text generated using T5 and OPUS-MT model with and with single-bit errors in the parameters of the LLM. The T5 LLM used the CNN Daily Mail dataset for summarization and OPUS-MT used the IWSLT2017 dataset for Chinese-to-English translation.

Files:

• {cnn/iwslt2017}_input_text.txt: Input text, that is, text to summarize (cnn and T5) or Chinese text to translate (iwslt2017 and OPUS-MT).  For each dataset in total there are number_input_texts.
• {cnn/iwslt2017}_output_reference.txt: Example of result expected for CNN (T5) and IWSLT2017 (OPUS-MT). For each dataset in total there are number_input_texts.
• {cnn/iwslt2017}_output_predict_fault_free: Example of predictions without single-bit errors. For each dataset in total there are number_input_texts.
• {cnn/iwslt2017}_output_predict_single_fi_bit_100times: Example of predictions with 100 different single-bit error. In each dataset in total there are 100*number input texts.

Paper

• Paper: Concurrent Linguistic Error Detection (CLED) for Large Language Models
• Cite:

@ARTICLE{11145323,
  author={Zhu, Jinhua and Conde, Javier and Gao, Zhen and Reviriego, Pedro and Liu, Shanshan and Lombardi, Fabrizio},
  journal={IEEE Transactions on Computers}, 
  title={Concurrent Linguistic Error Detection (CLED): a New Methodology for Error Detection in Large Language Models}, 
  year={2025},
  volume={},
  number={},
  pages={1-14},
  keywords={Protection;Feature extraction;Machine learning;Neural networks;Linguistics;Computational modeling;Electronic mail;Transformers;Large language models;Hardware;LLMs;soft errors;concurrent error detection;T5;OPUS-MT},
  doi={10.1109/TC.2025.3603682}}