Textual backdoor attacks are a kind of practical threat to NLP systems. By injecting a backdoor in the training phase, the adversary could control model predictions via predefined triggers. As various attack and defense models have been proposed, it is of great significance to perform rigorous evaluations. However, we highlight two issues in previous backdoor learning evaluations: (1) The differences between real-world scenarios (e.g. releasing poisoned datasets or models) are neglected, and we argue that each scenario has its own constraints and concerns, thus requires specific evaluation protocols; (2) The evaluation metrics only consider whether the attacks could flip the models' predictions on poisoned samples and retain performances on benign samples, but ignore that poisoned samples should also be stealthy and semantic-preserving. To address these issues, we categorize existing works into three practical scenarios in which attackers release datasets, pre-trained models, and fine-tuned models respectively, then discuss their unique evaluation methodologies. On metrics, to completely evaluate poisoned samples, we use grammar error increase and perplexity difference for stealthiness, along with text similarity for validity. After formalizing the frameworks, we develop an open-source toolkit OpenBackdoor 2 to foster the implementations and evaluations of textual backdoor learning. With this toolkit, we perform extensive experiments to benchmark attack and defense models under the suggested paradigm. To facilitate the underexplored defenses against poisoned datasets, we further propose CUBE, a simple yet strong clustering-based defense baseline. We hope that our frameworks and benchmarks could serve as the cornerstones for future model development and evaluations. * Equal contribution 2 https://github.com/thunlp/OpenBackdoor Preprint. Under review.

Pre-trained language models (PLMs) achieve remarkable performance on many downstream tasks, but may fail in giving reliable estimates of their predictive uncertainty. However, well-calibrated PLMs are essential for decision makers in high-risk applications since their confidence scores reflect the true probability of the outcomes. Given the lack of a comprehensive understanding of PLMs calibration, we take a close look into this new research problem, aiming to answer two questions: (1) Do PLMs learn to become calibrated in the training process? (2) How effective are existing calibration methods? For the first question, we conduct fine-grained control experiments to study the dynamic change in PLMs' calibration performance in training. We consider six factors as control variables, including dataset difficulty, available training samples, training steps, the number of tunable parameters, model scale, and pretraining. In experiments, we observe a consistent change in calibration performance across six factors. We find that PLMs don't learn to become calibrated in training, evidenced by the continual increase in confidence, no matter the predictions are correct or not. We highlight that our finding presents some contradiction with two established conclusions: (a) Larger PLMs are more calibrated; (b) Pretraining improves model calibration. Next, we study the effectiveness of existing calibration methods in mitigating the overconfidence issue, in both in-distribution and various out-of-distribution settings. Besides unlearnable calibration methods (e.g., label smoothing), we adapt two recently proposed learnable methods that directly collect data to train models to have reasonable confidence estimations. Also, we propose extended learnable methods based on existing ones to further improve or maintain PLMs calibration without sacrificing the original task performance. Experimental results show that learnable methods significantly reduce PLMs' confidence in wrong predictions, and our methods exhibit superior performance compared with previous methods. All the code and data can be obtained at https://github.com/lifan-yuan/PLMCalibration.

No abstract