Metraipolis › archive › dc_abc3c3507ddd
BadEngram: Backdoor Attack on Gated Memory Components in LLMs — arXiv:2609.13478
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.13478
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T23:53:34.298Z
- SHA-256 of the archived text
2ba9055b6a7b28dfc87d5e3e04f09d44189a7cd1221cb9db3306b604d4751322- Stable URL of this page
https://metraipolis.com/documento/dc_abc3c3507ddd
How to cite
BadEngram: Backdoor Attack on Gated Memory Components in LLMs — arXiv:2609.13478 — arXiv (en), https://arxiv.org/abs/2609.13478 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_abc3c3507ddd (Metraipolis archive dc_abc3c3507ddd, sha256 2ba9055b6a7b28dfc87d5e3e04f09d44189a7cd1221cb9db3306b604d4751322)
BibTeX
@misc{dcabc3c3507ddd,
title = {BadEngram: Backdoor Attack on Gated Memory Components in LLMs — arXiv:2609.13478},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.13478},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_abc3c3507ddd; sha256 2ba9055b6a7b28dfc87d5e3e04f09d44189a7cd1221cb9db3306b604d4751322},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.13478. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
To expand open-weight models' capacity without proportionally increasing computation, recent architectures incorporate gated parametric memory that retrieves learned values and injects them into intermediate representations. One representative design is Engram, which combines deterministic n-gram lookup with context-dependent gating over large learned memory tables. Despite these efficiency benefits, such modules create a distinct attack surface: their parameters can be modified independently of the backbone while directly shaping its computation. We introduce BadEngram, a post-training attack that exploits this surface to implant persistent, trigger-dependent behavior while leaving conventional backbone weights and the execution graph unchanged. We first establish the attack's feasibility and causally characterize its mechanism in a controlled Engram model, where BadEngram achieves 96.6% ASR on triggered inputs while limiting false activation on matched trigger-free inputs to 0.1% and preserving 99.6% clean accuracy. Replacing the retrieved memory values with their clean counterparts or closing the memory gates reduces ASR to at most 0.32%, confirming that the backdoor is expressed through the gated-memory pathway. We then test whether this vulnerability extends to production scale in Qwen3.8-Flash-Next's native Per-Layer Embedding subsystem. Using independently trained checkpoints for the two benchmarks, BadEngram achieves 47.8% ASR on HarmBench and 64.0% on AdvBench, while dormant-condition ASR remains 0.9% and 0.0%, respectively. These results identify native gated-memory parameters as a security-critical part of the model whose integrity cannot be inferred from an unchanged backbone. Tema degli esploratori: fiducia digitale e crittografia (retro-catalogo arXiv: annunci del 2026-09-15 in cs.CR). Categorie arXiv: Cryptography and Security (cs.CR). Pagina del lavoro: https://arxiv.org/abs/2609.13478 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-15T23:20:42.630Z dalla pagina https://arxiv.org/catchup/cs.CR/2026-09-15?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).