{"slug": "groundhog-bit-flip-attack-seeding-infinite-generation-loops-in-mixture-of-llms", "title": "Groundhog Bit-Flip Attack: Seeding Infinite Generation Loops in Mixture-of-Experts LLMs through Bit Flips", "summary": "Researchers introduced the Groundhog Bit-Flip Attack (GBFA), the first bit-flip-based Denial-of-Wallet availability attack against Mixture-of-Experts (MoE) large language models, which inflates decoding token usage by an average of 5912% across four real-world MoE-based LLMs by deactivating fewer than 4 experts on average. The attack targets routing-layer bits to extend output length while preserving semantic fidelity, exposing a robustness vulnerability in MoE architectures.", "body_md": "arXiv:2608.25276v1 Announce Type: new\nAbstract: Mixture-of-Experts (MoE) architectures enable scalable and efficient large language models (LLMs) by selectively activating expert sub-networks through a routing mechanism. However, this adaptive design introduces a new attack surface: specific experts become disproportionately correlated with certain tokens (e.g., end-of-sequence), allowing adversaries to manipulate model behavior via lightweight perturbations. In this work, we present \\textbf{Groundhog Bit-Flip Attack (GBFA)}, the first bit-flip-based \\textit{ Denial-of-Wallet availability attack} against MoE-based LLMs. By identifying and flipping routing-layer bits associated with related expert activations, we demonstrate that GBFA substantially extends the decoding token usage across three different LLM modes: conversational, reasoning, and agentic tasks, while largely preserving semantic fidelity. Across four main real-world MoE-based LLMs, manually deactivating on average fewer than \\textbf{4 experts} drives average output inflation to $\\mathbf{5912\\%}$, with the majority of test samples reaching max tokens. These results reveal a robustness vulnerability of MoE architectures to bit flip, and highlight the potential of GBFA as an availability attack against LLMs.", "url": "https://wpnews.pro/news/groundhog-bit-flip-attack-seeding-infinite-generation-loops-in-mixture-of-llms", "canonical_source": "https://arxiv.org/abs/2608.25276", "published_at": "2026-08-27 04:00:00+00:00", "updated_at": "2026-08-27 04:20:32.204762+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-safety", "ai-research"], "entities": ["Groundhog Bit-Flip Attack (GBFA)", "Mixture-of-Experts (MoE)"], "alternates": {"html": "https://wpnews.pro/news/groundhog-bit-flip-attack-seeding-infinite-generation-loops-in-mixture-of-llms", "markdown": "https://wpnews.pro/news/groundhog-bit-flip-attack-seeding-infinite-generation-loops-in-mixture-of-llms.md", "text": "https://wpnews.pro/news/groundhog-bit-flip-attack-seeding-infinite-generation-loops-in-mixture-of-llms.txt", "jsonld": "https://wpnews.pro/news/groundhog-bit-flip-attack-seeding-infinite-generation-loops-in-mixture-of-llms.jsonld"}}