# IMBERT: Making BERT Immune to Insertion-based Backdoor Attacks

arXiv: https://arxiv.org/abs/2305.16503 · HTML: https://mlsanity.com/paper/2305.16503/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- Placed in Backdoor and data poisoning attacks, one of 12 directions in LLM safety and security.
- Builds on 31 of its 48 references that are in the map.
- First posted on arXiv 25 May 2023.

## Placed in

- LLM safety and security / [Backdoor and data poisoning attacks](https://mlsanity.com/safety/backdoor-and-data-poisoning-attacks/)

## Builds on (most cited first, 30 of 31)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding](https://mlsanity.com/paper/1804.07461/) (2018)
- [Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps](https://mlsanity.com/paper/1312.6034/) (2013)
- [Character-level Convolutional Networks for Text Classification](https://mlsanity.com/paper/1509.01626/) (2015)
- [BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chain](https://mlsanity.com/paper/1708.06733/) (2017)
- [Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoning](https://mlsanity.com/paper/1712.05526/) (2017)
- [Attention is not not Explanation](https://mlsanity.com/paper/1908.04626/) (2019)
- [Is Attention Interpretable?](https://mlsanity.com/paper/1906.03731/) (2019)
- [Hidden Trigger Backdoor Attacks](https://mlsanity.com/paper/1910.00033/) (2019)
- [Reflection Backdoor: A Natural Backdoor Attack on Deep Neural Networks](https://mlsanity.com/paper/2007.02343/) (2020)
- [Weight Poisoning Attacks on Pre-trained Models](https://mlsanity.com/paper/2004.06660/) (2020)
- [Predicting the Type and Target of Offensive Posts in Social Media](https://mlsanity.com/paper/1902.09666/) (2019)
- [ONION: A Simple and Effective Defense Against Textual Backdoor Attacks](https://mlsanity.com/paper/2011.10369/) (2020)
- [A backdoor attack against LSTM-based text classification systems](https://mlsanity.com/paper/1905.12457/) (2019)
- [Hidden Killer: Invisible Textual Backdoor Attacks with Syntactic Trigger](https://mlsanity.com/paper/2105.12400/) (2021)
- [Clean-Label Backdoor Attacks on Video Recognition Models](https://mlsanity.com/paper/2003.03030/) (2020)
- [Learned Token Pruning for Transformers](https://mlsanity.com/paper/2107.00910/) (2021)
- [Backdoor Embedding in Convolutional Neural Network Models via Invisible Perturbation](https://mlsanity.com/paper/1808.10307/) (2018)
- [Turn the Combination Lock: Learnable Textual Backdoor Attacks via Word Substitution](https://mlsanity.com/paper/2106.06361/) (2021)
- [BadPre: Task-agnostic Backdoor Attacks to Pre-trained NLP Foundation Models](https://mlsanity.com/paper/2110.02467/) (2021)
- [AllenNLP Interpret: A Framework for Explaining Predictions of NLP Models](https://mlsanity.com/paper/1909.09251/) (2019)
- [Attention Interpretability Across NLP Tasks](https://mlsanity.com/paper/1909.11218/) (2019)
- [Backdooring Convolutional Neural Networks via Targeted Weight Perturbations](https://mlsanity.com/paper/1812.03128/) (2018)
- [Putting words into the system's mouth: A targeted attack on neural machine translation using monolingual data poisoning](https://mlsanity.com/paper/2107.05243/) (2021)
- [Mitigating Backdoor Poisoning Attacks through the Lens of Spurious Correlation](https://mlsanity.com/paper/2305.11596/) (2023)
- [Magic Pyramid: Accelerating Inference with Early Exiting and Token Pruning](https://mlsanity.com/paper/2111.00230/) (2021)

## Built on by (most cited first, 15 of 15)

- [Use of LLMs for Illicit Purposes: Threats, Prevention Measures, and Vulnerabilities](https://mlsanity.com/paper/2308.12833/) (2023)
- [Composite Backdoor Attacks Against Large Language Models](https://mlsanity.com/paper/2310.07676/) (2023)
- [Backdoor Attacks and Countermeasures in Natural Language Processing Models: A Comprehensive Security Review](https://mlsanity.com/paper/2309.06055/) (2023)
- [Here's a Free Lunch: Sanitizing Backdoored Models with Model Merge](https://mlsanity.com/paper/2402.19334/) (2024)
- [BadActs: A Universal Backdoor Defense in the Activation Space](https://mlsanity.com/paper/2405.11227/) (2024)
- [TuBA: Cross-Lingual Transferability of Backdoor Attacks in LLMs with Instruction Tuning](https://mlsanity.com/paper/2404.19597/) (2024)
- [Turning Generative Models Degenerate: The Power of Data Poisoning Attacks](https://mlsanity.com/paper/2407.12281/) (2024)
- [Concept-Level Explainability for Auditing & Steering LLM Responses](https://mlsanity.com/paper/2505.07610/) (2025)
- [SEEP: Training Dynamics Grounds Latent Representation Search for Mitigating Backdoor Poisoning Attacks](https://mlsanity.com/paper/2405.11575/) (2024)
- [Detecting Stealthy Backdoor Samples based on Intra-class Distance for Large Language Models](https://mlsanity.com/paper/2505.23015/) (2025)
- [NLP Security and Ethics, in the Wild](https://mlsanity.com/paper/2504.06669/) (2025)
- [SLIP: Soft Label Mechanism and Key-Extraction-Guided CoT-based Defense Against Instruction Backdoor in APIs](https://mlsanity.com/paper/2508.06153/) (2025)
- [SteganoBackdoor: Evading Data-Poisoning Defenses via Steganographic Backdoors](https://mlsanity.com/paper/2511.14301/) (2025)
- [Backdoor Samples Detection Based on Perturbation Discrepancy Consistency in Pre-trained Language Models](https://mlsanity.com/paper/2509.05318/) (2025)
- [An Empirical Study of Output-to-Input Loops for Black-Box Backdoor Detection in Fine-Tuned Open-Weight LLMs](https://mlsanity.com/paper/2608.11348/) (2026)

## Abstract (arXiv)

Backdoor attacks are an insidious security threat against machine learning models. Adversaries can manipulate the predictions of compromised models by inserting triggers into the training phase. Various backdoor attacks have been devised which can achieve nearly perfect attack success without affecting model predictions for clean inputs. Means of mitigating such vulnerabilities are underdeveloped, especially in natural language processing. To fill this gap, we introduce IMBERT, which uses either gradients or self-attention scores derived from victim models to self-defend against backdoor attacks at inference time. Our empirical studies demonstrate that IMBERT can effectively identify up to 98.5% of inserted triggers. Thus, it significantly reduces the attack success rate while attaining competitive accuracy on the clean dataset across widespread insertion-based attacks compared to two baselines. Finally, we show that our approach is model-agnostic, and can be easily ported to several pre-trained transformer models.