arXiv Computation and Language September 25, 2026 By Wissam Antoun, Francis Kulumba, Th\'eo Lasnier, Beno\^it Sagot, Djam\'e Seddah

LLM Forensics: Where Do Backdoors Hide? Localizing and Controlling Trigger Mechanisms with Sparse Autoencoders

Read the original on arXiv Computation and Language →

The paper investigates how trigger-based backdoors function in large language models by using sparse autoencoders (SAEs) to identify feature directions across layers and transformer components. In a controlled experiment, 1B and 8B models were made to continue English prompts in French or German when presented with fixed trigger sequences. The study finds that different SAE feature directions correspond to trigger detection, residual-stream propagation, and language tracking, with residual-stream features being most effective for controlling the backdoor behavior.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computation and Language.

llms

arXiv AI

Jun 9

Shared Latent Structures Enable Unified Backdoor Detection and Mitigation in LLMs

arXiv:2606. 07963v1 Announce Type: new Abstract: Backdoor attacks in large language models (LLMs) are often treated as isolated trigger-response failures, motivating defenses tailored to specific triggers or behaviors.

By Omar Mahmoud, Aly M. Kassem, Thommen George Karimpanal, Buddhika Laknath Semage, Negar Rostamzadeh, Golnoosh Farnadi, Santu Rana

llms robotics fine-tuning benchmarks safety

arXiv Computation and Language

Aug 25

Forgetting to Forget: Attention Sink as A Gateway for Backdooring LLM Unlearning

arXiv:2510.17021v2 Announce Type: replace-cross Abstract: Large language model (LLM) unlearning is a key approach for removing undesired data, knowledge, or behaviors from pretrained models while ret...

By Bingqi Shang, Yiwei Chen, Yihua Zhang, Bingquan Shen, Sijia Liu

llms

arXiv AI

Sep 2

Detecting Hidden Behaviors in LLMs via Activation-matched Finetuning

arXiv:2609.00351v1 Announce Type: cross Abstract: Large language models can hide hidden behaviors that activate only under narrow conditions, such as backdoor triggers, sleeper-agent deployment cues,...

By Robin Haselhorst, Lucie Flek, Florian Mai

llms agents fine-tuning

Hugging Face Trending Papers

Jun 2

Backdoor Unlearning Generalization: A Path Toward the Removal of Unknown Triggers in LLMs

Backdoor attacks in Large Language Models (LLMs) are a growing security concern, where models can generate adversary-chosen content. Existing defenses target backdoors one at a time and typically require knowledge of the trigger, leaving the defender at a structural disadvantage when unknown backdoors may exist in a model.

llms safety

arXiv AI

Sep 7

Recurrence Is Not Enough: Causally Validating Multilingual SAE Translation Features in Gemma 2 and 3

The paper investigates whether sparse autoencoder (SAE) features that recur across different language settings in Gemma 2 and Gemma 3 actually have consistent causal effects on translation performance. By reproducing Wu et al.’s discovery method and extending it to multilingual prompts, the authors find over 20 frequently activating features, yet causal validation reveals that almost all have negligible or inconsistent impacts. Only one feature—Gemma 2’s (L10, 5717) and Gemma 3’s (L20, 2456)—consistently improves COMET scores when amplified and worsens them when ablated across 23 language settings, indicating a language‑agnostic translation‑initiation direction.

By Giang Son Nguyen, Nhi Ngoc-Yen Nguyen, Wray Buntine, Dung D. Le

arXiv Machine Learning

Sep 10

LLM Layers Immediately Correct Each Other

arXiv:2609.07876v1 Announce Type: cross Abstract: Recent methods in language model interpretability employ techniques such as sparse autoencoders to decompose residual stream contributions into linea...

By Arjun Patrawala, Jiahai Feng, Erik Jones, Jacob Steinhardt

llms safety

One email a morning, machine-written

Related stories

Shared Latent Structures Enable Unified Backdoor Detection and Mitigation in LLMs

Forgetting to Forget: Attention Sink as A Gateway for Backdooring LLM Unlearning

Detecting Hidden Behaviors in LLMs via Activation-matched Finetuning

Backdoor Unlearning Generalization: A Path Toward the Removal of Unknown Triggers in LLMs

Recurrence Is Not Enough: Causally Validating Multilingual SAE Translation Features in Gemma 2 and 3

LLM Layers Immediately Correct Each Other