Breaking Safety at the Token Boundary: How BPE Tokenization Creates Exploitable Gaps in LLM Alignment

Summary

arXiv:2607.01239v1 Announce Type: new Abstract: Character-level perturbations bypass safety alignment in modern LLMs despite leaving prompts human-readable. We identify and test a central structural mechanism: BPE tokenization fragments safety-critical words into sub-word pieces, and the three public alignment datasets we surveyed contain no intentionally fragmented inputs. The mechanism is a chain, tested end-to-end on five model families (Qwen-3-4B, Qwen-2.5-7B, Gemma-3-4B, Llama-3.1-8B, Mistr

SpinGraph analysis pending — check back after processing.

• Can Language Models Actually Retrieve In-Context? Drowning in Documents at Million Token Scale
• Parameter Golf: What Really Works?
• From Monolingual to Multilingual: Evaluating Mamba for ASR in South African Languages
• Comparing Architectures for Supervised Political Scaling
• Grounded Optimization: A Layered Engineering Framework for Reducing LLM Hallucination in Automated Personal Document Rewriting
• FaithMed: Training LLMs For Faithful Evidence-Based Medical Reasoning