Adaptive Adversarial Prompt Detection in Large Language Models via Quantum-Resistant Cryptography and Spectral–Spatial Wave Networks
Bibliographic record
Abstract
The rapid growth of large language models (LLMs) has intensified concerns over adversarial prompts, which can manipulate model outputs, bypass safety constraints, or cause unintended behaviors. Traditional prompt filtering systems rely on keyword-based detection or basic semantic similarity checks, which often fail to identify obfuscated, rephrased, or structurally disguised malicious inputs. Furthermore, many existing solutions store prompts in plain text for analysis, creating a significant security vulnerability. To address these challenges, a secure and adaptive adversarial prompt detection framework is introduced, integrating post-quantum cryptography, advanced semantic–behavioral analysis, and metaheuristic optimization. In this approach, incoming prompts are immediately encrypted using a Quantum-Resistant Cryptography algorithm before any storage, ensuring confidentiality even under future quantum attacks. A normalization-based preprocessing stage removes obfuscations, malicious syntax, and hidden tokens, followed by a Hierarchical Attention Transformer for combined semantic and behavioral analysis, capturing both local syntactic cues and global semantic dependencies. The extracted features are processed through a self-attention-based Spectral–Spatial Wave Network(Self-SSWN), whose parameters are optimized using the Migrating Walrus Algorithm to enhance detection accuracy. The optimized deep learning classifier assigns confidence-based threat levels—benign, potentially adversarial, or confirmed adversarial—while a continuous feedback loop updates the model with newly detected attack patterns.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.008 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.001 | 0.003 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".