This dataset is part of the RobustFT project: Robust Supervised Fine-tuning for Large Language Models under Noisy Response. The dataset contains various test cases with different noise ratios for training and evaluating robust fine-tuning approaches.
Our paper:
https://huggingface.co/papers/2412.14922RobustFT/
├── arc/
│ │── noisy30.csv
│ │── noisy50.csv
│ │── noisy70.csv
│ ├── labeled.csv
│ └── test.csv
├── drop/
│ │── noisy30.csv
│… See the full description on the dataset page:
https://huggingface.co/datasets/luojunyu/RobustFT.
