Research map / Vision-language and multimodal / Multimodal LLMs
DecepGPT: Schema-Driven Deception Detection with Multicultural Datasets and Robust Multimodal Learning
JIAJIAN HUANG, Dongliang Zhu, Zitong Yu, Hui Ma, Jiayu Zhang, Chunmei Zhu, Xiaochun Cao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add SICS adapter (ablation) | DecepGPT base · DOLOs (in-domain) | accuracy (%): 67.24 -> 72.68 | Improved |
| add DMC regularizer (ablation) | DecepGPT base · DOLOs (in-domain) | accuracy (%): 67.24 -> 68.77 | Improved |
| add DMC on top of SICS (ablation) | DecepGPT + SICS · DOLOs (in-domain) | accuracy (%): 72.68 -> 73.23 | Improved |
| add SICS adapter (ablation) | DecepGPT base · Bag-of-Lies (in-domain) | accuracy (%): 57.69 -> 62.50 | Improved |
| add DMC regularizer (ablation) | DecepGPT base · MU3D (in-domain) | accuracy (%): 53.75 -> 55.00 | Improved |
| add SICS adapter (ablation) | DecepGPT base · MU3D (in-domain) | accuracy (%): 53.75 -> 60.15 | Improved |
| add SICS adapter (ablation) | DecepGPT base · Cross-domain average | accuracy (%): 51.94 -> 58.04 | Improved |
| add DMC regularizer (ablation) | DecepGPT base · Cross-domain average | accuracy (%): 51.94 -> 54.17 | Improved |
| add global prior b_global to SICS (ablation) | SICS adapter · DOLOs (in-domain) | accuracy (%): 71.42 -> 73.23 | Improved |
| add gating to SICS (ablation) | SICS adapter · DOLOs (in-domain) | accuracy (%): 71.85 -> 73.23 | Improved |
| add polarity-aware adjustment to SICS (ablation) | SICS adapter · DOLOs (in-domain) | accuracy (%): 72.06 -> 73.23 | Improved |
| use LLM backbone instead of MLP (ablation) | DecepGPT · DOLOs (in-domain) | accuracy (%): 58.42 -> 73.23 | Improved |
Closest papers by meaning
- Debate with Images: Detecting Deceptive Behaviors in Multimodal Large Language Models (ICML 2026 · similarity 0.88)
- [Re] Boosting the Visual Interpretability of CLIP via Adversarial Fine-Tuning (NeurIPS 2026 · similarity 0.86)
- XFeat Revisited: Reproducibility and Evaluation of a Lightweight Image Matcher (NeurIPS 2026 · similarity 0.86)
- The Coherence Trap: When MLLM-Crafted Narratives Exploit Manipulated Visual Contexts (CVPR 2026 · similarity 0.85)
- How to Catch an AI Liar: Lie Detection in Black-Box LLMs by Asking Unrelated Questions (ICLR 2024 · similarity 0.85)
- Liars' Bench: Evaluating Lie Detectors for Language Models (NeurIPS 2026 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
