Adaptive LLM-Driven Phishing Defense Using Real-Time Psychological Cue Detection
Bibliographic record
Abstract
Phishing attacks remain one of the most prevalent and damaging cybersecurity threats, with attackers increasingly employing sophisticated psychological manipulation techniques to deceive victims. This paper introduces an innovative adaptive defense system that leverages Large Language Models (LLMs) and real-time psychological cue detection to enhance phishing threat detection and mitigation. By integrating psychological cues, such as urgency, fear, and social influence, detected from email and web communication patterns, the system dynamically adapts its defense strategies in real-time. The proposed framework continuously analyzes both the content and context of messages, using LLMs to assess linguistic features, detect inconsistencies, and identify manipulative tactics employed by cybercriminals. A psychological cue-based risk model is developed, enabling the system to predict the likelihood of phishing attacks based on the emotional tone and behavioral triggers embedded within the communication. The effectiveness of the approach is demonstrated through experimental results, showing a significant improvement in phishing detection accuracy and reduced false positive rates compared to traditional methods. This adaptive, AI-driven model provides a robust solution for defending against the evolving landscape of phishing attacks, offering both proactive and reactive capabilities in real-world environments.
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 | 0.001 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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".