The Tokens Remember: When Tokenization Bypasses Knowledge Editing and Unlearning

arXiv:2609.29045 · cs.CR, cs.AI · Submitted 2026-09-24 · Read on arXiv

cs.CR, cs.AI

Submitted: 2026-09-24

Updated: 2026-09-24

Terminology

• Tokenization — 254× in this paper · explained in 2 episodes
• Entanglement — 116× in this paper · explained in 4 episodes
• AUC — 79× in this paper · explained in 2 episodes
• Fine-Tuning — 18× in this paper · explained in 5 episodes
• Calibration — 17× in this paper · explained in 7 episodes
• Machine Unlearning — 16× in this paper · explained in 2 episodes
• Robustness — 16× in this paper · explained in 5 episodes
• Accuracy — 13× in this paper · explained in 1 episode
• Benchmark — 13× in this paper · explained in 2 episodes
• Alignment — 8× in this paper · explained in 1 episode
• F1 Score — 6× in this paper · explained in 3 episodes
• Decode — 5× in this paper · explained in 2 episodes
• Generalization — 4× in this paper · explained in 13 episodes
• RAG — 4× in this paper · explained in 3 episodes
• Transformer — 4× in this paper · explained in 2 episodes
• Core — 3× in this paper · explained in 2 episodes
• Foundation Model — 3× in this paper · explained in 3 episodes
• Foundation Models — 3× in this paper · explained in 7 episodes
• Adaptation — 2× in this paper · explained in 1 episode
• Adaptive Sampling — 2× in this paper · explained in 1 episode
• Expected Calibration Error — 2× in this paper
• Generative AI — 2× in this paper · explained in 2 episodes
• In-Context Learning — 2× in this paper · explained in 4 episodes
• Maximum Calibration Error — 2× in this paper
• Regularization — 2× in this paper · explained in 3 episodes
• Accountability — 1× in this paper · explained in 2 episodes
• Actor — 1× in this paper · explained in 3 episodes
• Benchmarking — 1× in this paper · explained in 4 episodes
• Byte-Pair Encoding — 1× in this paper
• Commentary — 1× in this paper · explained in 2 episodes
• Cosine Similarity — 1× in this paper · explained in 3 episodes
• Domain Adaptation — 1× in this paper · explained in 5 episodes
• Generative Recommendation — 1× in this paper · explained in 2 episodes
• Hallucination — 1× in this paper · explained in 4 episodes
• Interpretability — 1× in this paper · explained in 15 episodes
• Knowledge Graph — 1× in this paper · explained in 7 episodes
• Knowledge graphs — 1× in this paper · explained in 2 episodes
• Large Language Models (LLMs) — 1× in this paper · explained in 6 episodes
• Mechanistic interpretability — 1× in this paper · explained in 2 episodes
• Quantization — 1× in this paper · explained in 12 episodes
• Retrieval-Augmented Generation (RAG) — 1× in this paper · explained in 16 episodes

Sources

• DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence
• GLM-5: from Vibe Coding to Agentic Engineering
• Olmo 3
• Shieldstral
• Llama-3.1-FoundationAI-SecurityLLM-Base-8B Technical Report
• Small, Free, and Effective: Orchestrating Open-Weight Small Language Models to Outperform Single LLM for Malware Analysis
• Paladin: Defending LLM-enabled Phishing Emails with a New Trigger-Tag Paradigm
• Mark Your LLM: Detecting the Misuse of Open-Source Large Language Models via Watermarking
• The Llama 3 Herd of Models
• Qwen3 Technical Report
• Ministral 3
• Gemma 4 Technical Report
• GROM: Gradient-Free Rapid One-Shot Machine Unlearning
• Industrialized Deception: The Collateral Effects of LLM-Generated Misinformation on Digital Ecosystems
• Reverse-Engineering Model Editing on Language Models
• From Large to Super-Tiny: End-to-End Optimization for Cost-Efficient LLMs
• FLEKE: Federated Locate-then-Edit Knowledge Editing
• AnyEdit: Edit Any Knowledge Encoded in Language Models
• Lifelong Knowledge Editing requires Better Regularization
• GenRecEdit: Adapting Model Editing for Generative Recommendation with Cold-Start Items

Related papers

• SoK: AI-Augmented Binary Reversing
• Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
• Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
• Efficient Fuzzy PSI under One-Sided Assumptions
• Sealing the Audit-Runtime Gap for LLM Skills
• Token Composition: A Graph Based on EVM Logs