Conspiracy theories on social media pose significantsocietal risks, making it essential todetect both conspiracy-related content and thetextual spans that serve as conspiracy markers.In this work, we propose two effective methodsto address these challenges. For markerextraction, we develop a span-based slidingwindow framework that improves efficiencyand accuracy by focusing on localized context.In addition, inspired by the distinctive emotionalpatterns in conspiracy texts, we designa dynamic gating mechanism to integrate emotionaland semantic representations. We evaluateour methods on the SemEval 2026 Task 10,where our team (dangphuduy) achieved competitiveresults, ranking 4th in Task 1 (SpanExtraction) and 3rd in Task 2 (Conspiracy Detection).Experimental results demonstrate thatboth proposed methods significantly enhancemodel performance.

The Opioid Industry Documents Archive (OIDA) provides extensive internal corporate records that offer valuable insight into the drivers of the opioid crisis, yet its use in systematic analysis of corporate strategy remains limited. In this study, we propose an NLP-based framework to analyze strategic behavior in large-scale litigation archives, combining relevance filtering and topic modeling with large language model (LLM)-assisted interpretation. Applied to documents from Insys Therapeutics and Mallinckrodt Pharmaceuticals, our approach uncovers systematic differences in corporate strategies and organizational priorities. These results highlight the potential of integrating representation learning and LLMs for large-scale analysis in public health and corporate accountability research.