Metraipolis › archive › dc_993d38c70697
Evaluating and Improving the Robustness of Large Language Models to Input Sequence Variations — arXiv:2610.02432
- Source
- arXiv
- Original
- https://arxiv.org/abs/2610.02432
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-05T22:51:54.804Z
- SHA-256 of the archived text
18335932600847629a4b2856b971ed6adbda50b14596915b2c0d768146198ab9- Stable URL of this page
https://metraipolis.com/documento/dc_993d38c70697
How to cite
Evaluating and Improving the Robustness of Large Language Models to Input Sequence Variations — arXiv:2610.02432 — arXiv, https://arxiv.org/abs/2610.02432 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_993d38c70697 (Metraipolis archive dc_993d38c70697, sha256 18335932600847629a4b2856b971ed6adbda50b14596915b2c0d768146198ab9)
BibTeX
@misc{dc993d38c70697,
title = {Evaluating and Improving the Robustness of Large Language Models to Input Sequence Variations — arXiv:2610.02432},
howpublished = {arXiv},
url = {https://arxiv.org/abs/2610.02432},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_993d38c70697; sha256 18335932600847629a4b2856b971ed6adbda50b14596915b2c0d768146198ab9},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.02432. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Large language models (LLMs) in production systems face prompt injections, trojans (backdoors), and manipulation of automatic quality metrics. This thesis develops models, methods, and algorithms for evaluating and improving LLM robustness to adversarial input sequence variations. We propose R_stab(f), a generative robustness metric based on the Jensen-Shannon divergence between per-step output distributions under small input perturbations. For localized attacks we prove V(h) <= 1 - R_class(h), where R_class(h) is the probability that a decision operator h keeps its decision under small perturbations. For non-localized attacks we propose a calibrated empirical model. For LLM-as-a-Judge systems we develop ASA, an adaptive evolutionary black-box attack that reaches an attack success rate (ASR) of up to 73.8%, with transfer between open models up to 62.6%. On Trojan Detection Challenge 2023 data (Pythia-1.4B), surrogate triggers reach REASR ~0.99 while recall of the true triggers is ~0.17 against a baseline of ~0.14. On SaTML CTF 2024 we systematize four classes of bypasses of multi-layer defenses, which reduce the ASR from 90% to 15-25%. Committees of 5-7 heterogeneous models reduce the ASR for Gemma-3-4B by 47-55 percentage points, to 19.3% with 7 models. For agentic systems based on the Model Context Protocol (MCP), we propose AttestMCP, which attests tool calls with HMAC-protected packets at under 0.1 ms per call, and the Commit Boundary isolation pattern. On the MCPBench benchmark of 847 scenarios they reduce the average ASR from 53.7% to 12.4%. The methods are implemented in the JudgeGuard and TrojanArmor software suites and the MCPSec module. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno). Categorie arXiv: Cryptography and Security (cs.CR) ; Computation and Language (cs.CL); Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.02432 — 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 2026-10-05T22:46:10.608Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).