CUET320 at SemEval-2026 Task 10: Few-Shot Large Language Models for Psycholinguistic Marker Extraction and Conspiracy Detection

2026-06-25 · Source: Paper Index on ACL Anthology · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Expert, quick

Summary

CUET320's system at SemEval-2026 Task 10 explored few-shot large language models (LLMs) for psycholinguistic marker extraction and binary conspiracy detection from Reddit submissions. The team adopted a training-free few-shot prompting approach, utilizing various instruction-tuned LLMs across different few-shot settings (k in {0,1,5,10,15,20}). Their prompting strategy incorporated psychology-informed instructions to guide models in identifying conspiracy-related signals. The system achieved an F1 score of 0.21 for marker extraction, ranking 16th out of 30 teams in Subtask 1, and 0.81 for conspiracy detection, placing 36th out of 52 in Subtask 2, without task-specific fine-tuning. These results indicate that psycholinguistically grounded prompting can support interpretable conspiracy analysis, despite challenges in identifying implicit markers.

Key takeaway

For NLP engineers developing automated systems for content moderation, consider integrating psycholinguistically-informed few-shot LLM prompting. This approach offers an interpretable method for conspiracy detection without requiring extensive task-specific fine-tuning. While effective for explicit signals, you should anticipate challenges with implicit markers and potentially combine this with other techniques for comprehensive coverage of subtle conspiratorial language.

Key insights

Psychology-informed few-shot LLM prompting aids interpretable conspiracy detection, though implicit markers remain challenging.

Principles

• Psycholinguistic markers enhance interpretability in detection.
• Few-shot prompting enables domain generalization without fine-tuning.
• LLMs face difficulties identifying implicit conspiracy signals.

Method

Training-free few-shot prompting with instruction-tuned LLMs, varying k (0,1,5,10,15,20), and incorporating psychology-informed instructions for identifying conspiracy-related signals.

In practice

• Apply psychology-informed prompts for LLM guidance.
• Experiment with diverse few-shot settings for new domains.
• Prioritize explicit markers for higher detection accuracy.

Topics

• Conspiracy Detection
• Large Language Models
• Few-Shot Learning
• Psycholinguistics
• SemEval-2026
• Social Media Analysis

Best for: Research Scientist, AI Scientist, NLP Engineer

Related on AIssential

• mdok-style at SemEval-2026 Task 10: Finetuning LLMs for Conspiracy Detection — Data augmentation and self-training can effectively finetune LLMs for specialized text classification tasks with limited data.
• Supervision versus Demonstration-Based In-Context Learning for Multiword Expression Classification — In-context learning for Turkish LVCs is highly prompt-sensitive, with few-shot demonstrations improving LLM calibration to match supervised baselines.
• ttda704 at SemEval-2026 Task 6: Structured Chain-of-Thought Prompting for Political Evasion Detection — Structured Chain-of-Thought prompting significantly improves political evasion detection over fine-tuning large language models.
• Exposing biases, moods, personalities, and abstract concepts hidden in large language models — A new method can identify and manipulate abstract concepts hidden within large language models to enhance or minimize their expression.
• Detecting HIV-Related Stigma in Clinical Narratives Using Large Language Models — LLMs can effectively detect HIV-related stigma in clinical notes, with encoder models outperforming generative models in zero-shot settings.
• Do we really need to detect LLM-generated text? — Masked word prediction by LLMs can distinguish AI-generated text from human text without explicit training.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Paper Index on ACL Anthology.