Hidden State Poisoning Attacks against Mamba-based Language Models

arXiv:2601.01972 · cs.CL, cs.AI, cs.LG · Submitted 2026-01-05 · Read on arXiv

cs.CL, cs.AI, cs.LG

Submitted: 2026-01-05

Updated: 2026-09-01

Comments: 27 pages, 4 figures. Accepted to Findings of the Association for Computational Linguistics: EMNLP 2026

Code: https://github.com/TortueSagace/hispa

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Transformer — 32× in this paper · explained in 2 episodes
• Benchmark — 31× in this paper · explained in 2 episodes
• Robustness — 14× in this paper · explained in 5 episodes
• Accuracy — 9× in this paper · explained in 1 episode
• Core — 8× in this paper · explained in 2 episodes
• Prompt Injection Attack — 7× in this paper
• Adaptation — 4× in this paper · explained in 1 episode
• Interpretability — 4× in this paper · explained in 15 episodes
• Loss Function — 4× in this paper · explained in 2 episodes
• Actor — 3× in this paper · explained in 3 episodes
• Adversarial Robustness — 3× in this paper · explained in 3 episodes
• Calibration — 3× in this paper · explained in 7 episodes
• State Space Models (SSMs) — 3× in this paper · explained in 4 episodes
• Verification — 3× in this paper · explained in 2 episodes
• Hybrid Architectures — 2× in this paper · explained in 3 episodes
• Hybrid Models — 2× in this paper · explained in 2 episodes
• Prompt Engineering — 2× in this paper · explained in 4 episodes
• Transformer Architecture — 2× in this paper · explained in 2 episodes
• Adam — 1× in this paper · explained in 2 episodes
• Adversarial Attacks — 1× in this paper · explained in 1 episode
• Adversarial Training — 1× in this paper · explained in 1 episode
• Aggregation — 1× in this paper · explained in 1 episode
• Architecture-Specificity — 1× in this paper · explained in 1 episode
• Attention Mechanism — 1× in this paper · explained in 3 episodes
• Benchmarking — 1× in this paper · explained in 3 episodes
• Chain-of-Thought (CoT) — 1× in this paper · explained in 4 episodes
• Consistency — 1× in this paper · explained in 3 episodes
• Decode — 1× in this paper · explained in 2 episodes
• Distribution Shift — 1× in this paper · explained in 5 episodes
• Embeddings — 1× in this paper · explained in 2 episodes
• Fine-Tuning — 1× in this paper · explained in 5 episodes
• Generalization — 1× in this paper · explained in 13 episodes
• Large Language Models (LLMs) — 1× in this paper · explained in 6 episodes
• Multi-hop Reasoning — 1× in this paper · explained in 3 episodes
• Perplexity — 1× in this paper · explained in 3 episodes
• Regularization — 1× in this paper · explained in 3 episodes
• Reinforcement Learning — 1× in this paper · explained in 4 episodes
• Reinforcement Learning (RL) — 1× in this paper · explained in 25 episodes
• Transformer Models — 1× in this paper · explained in 2 episodes

Sources

• BlackMamba: Mixture of Experts for State-Space Models
• Nemotron-H: A Family of Accurate and Efficient Hybrid Mamba-Transformer Models
• NVIDIA Nemotron 3: Efficient and Open Intelligence
• Defending Against Prompt Injection With a Few DefensiveTokens
• Achilles' Heel of Mamba: Essential difficulties of the Mamba architecture demonstrated by synthetic data
• Mind the Confidence Gap: Overconfidence, Calibration, and Distractor Effects in Large Language Models
• Hymba: A Hybrid-head Architecture for Small Language Models
• Investigating the Indirect Object Identification circuit in Mamba
• The Pile: An 800GB Dataset of Diverse Text for Language Modeling
• Zamba: A Compact 7B SSM Hybrid Model
• Efficiently Modeling Long Sequences with Structured State Spaces
• RULER: What's the Real Context Size of Your Long-Context Language Models?
• Prompt Injection Attacks on LLM Generated Reviews of Scientific Publications
• Same Task, More Tokens: the Impact of Input Length on the Reasoning Performance of Large Language Models
• Jamba: A Hybrid Transformer-Mamba Language Model
• Hidden Prompts in Manuscripts Exploit AI-Assisted Peer Review
• Carbon Emissions and Large Neural Network Training
• Samba: Simple Hybrid State Space Models for Efficient Unlimited Context Language Modeling
• Recurrent Neural Networks (RNNs): A gentle Introduction and Overview
• Locating and Editing Factual Associations in Mamba

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering