This paper presents the K-NLPers system for SemEval-2026 Task 7: Everyday Knowledge Across Diverse Languages and Cultures. The task extends the BLEnD benchmark to evaluate cultural understanding of language models across more than 30 language-country pairs. Although Large Language Models (LLMs) achieve strong overall performance, they exhibit performance disparities across cultural contexts and tend to produce regionally biased responses. To address this limitation, we propose a continent-based multi-agent debate framework that leverages culture-specific performance differences instead of relying on a single model. For the Short Answer Question (SAQ) track, we employ three agents: a general-purpose model, a continent-specific model, and a country-level or culturally adjacent model. These agents engage in independent generation, mutual refinement, and final adjudication. For the Multiple-Choice Question (MCQ) track, we adopt a debate structure centered on high-performing general-purpose models due to the track’s simpler structure. Our system participated in all language-region pairs and achieved overall scores of 55.75 on SAQ and 88.32 on MCQ. Further analysis reveals that grouping the performance of various individual models by continent explains performance patterns more consistently than language-based grouping, highlighting the importance of cultural and historical context in model generalization.

Multilingual retrieval-augmented generation (mRAG) is often implemented within a fixed retrieval space, typically via query or document translation or multilingual embedding vector representations. However, this approach may be inadequate for culturally grounded queries, in which retrieval-condition misalignment may occur. Even strong retrievers and generators may struggle to produce culturally relevant answers when sourcing evidence from inappropriate linguistic or regional contexts. To this end, we introduce CORAL (COntext-aware Retrieval with Agentic Loop, an adaptive retrieval methodology for mRAG that enables iterative refinement of both the retrieval space (corpora) and the retrieval probe (query) based on the quality of the evidence. The overall process includes: (1) selecting corpora, (2) retrieving documents, (3) critiquing evidence for relevance and cultural alignment, and (4) checking sufficiency. If the retrieved documents are insufficient to answer the query correctly, the system (5) reselects corpora and rewrites the query. Across two cultural QA benchmarks, CORAL achieves up to a 3.58%p accuracy improvement on low-resource languages relative to the strongest baselines.

This paper presents automatic evaluation systems for assessing the pedagogical capabilities of LLM-based AI tutors. Drawing from a shared task, our systems specifically target four key dimensions of tutor responses: Mistake Identification, Mistake Location, Providing Guidance, and Actionability. These dimensions capture the educational quality of responses from multiple perspectives, including the ability to detect student mistakes, accurately identify error locations, provide effective instructional guidance, and offer actionable feedback. We propose GPT-4.1-based automatic evaluation systems, leveraging their strong capabilities in comprehending diverse linguistic expressions and complex conversational contexts to address the detailed evaluation criteria across these dimensions. Our systems were quantitatively evaluated based on the official criteria of each track. In the Mistake Location track, our evaluation systems achieved an Exact macro F1 score of 58.80% (ranked in the top 3), and in the Providing Guidance track, they achieved 56.06% (ranked in the top 5). While the systems showed mid-range performance in the remaining tracks, the overall results demonstrate that our proposed automatic evaluation systems can effectively assess the quality of tutor responses, highlighting their potential for evaluating AI tutor effectiveness.

In abstractive summarization, the challenge of producing concise and accurate summaries arises from the vast amount of information contained in the source document. Consequently, although Large Language Models (LLMs) can generate fluent text, they often introduce inaccuracies by hallucinating content not found in the original source. While supervised fine-tuning methods that maximize likelihood contribute to this issue, they do not consistently enhance the faithfulness of the summaries. Preference-based optimization methods, such as Direct Preference Optimization (DPO), can further refine the model to align with human preferences. However, these methods still heavily depend on costly human feedback. In this work, we introduce a novel and straightforward approach called Model-based Preference Optimization (MPO) to fine-tune LLMs for improved summarization abilities without any human feedback. By leveraging the model’s inherent summarization capabilities, we create a preference dataset that is fully generated by the model using different decoding strategies. Our experiments on standard summarization datasets and various metrics demonstrate that our proposed MPO significantly enhances the quality of generated summaries without relying on human feedback. The code is publicly available at https://github.com/cjaep/MPO.