{"slug": "llm-instruct-at-uzh-shared-task-2026-constraint-aware-retrieval-and-selective", "title": "LLM-INSTRUCT at UZH Shared Task 2026: Constraint-Aware Retrieval and Selective Debate for Paragraph-Level Argument Mining", "summary": "LLM-INSTRUCT, the winning system for the UZH Shared Task at ArgMining 2026, achieved 1st overall on the official leaderboard for paragraph-level argument mining in UN and UNESCO resolutions by using constraint-aware retrieval and selective debate. The system, which operates under a strict JSON schema with open-weight models up to 8B parameters, improved Task 1b Micro-F1 from 35.83% to 40.08% during development while maintaining an internal Task 2 score of 4.421. The team's code is publicly available on GitHub.", "body_md": "arXiv:2607.20430v1 Announce Type: new\nAbstract: We present LLM-INSTRUCT, the winning system for the UZH Shared Task at ArgMining 2026 on paragraph-level argument mining in UN and UNESCO resolutions. The task requires paragraph-type classification, prediction of a subset of 141 official tags, and directed relation prediction under a strict JSON schema setting using only open-weight models up to 8B parameters. We frame the task as constrained structured prediction. The system first narrows the candidate tag space with metadata-aware dense retrieval, then applies constrained decoding with per-dimension caps, escalates only uncertain cases to a three-agent debate branch, and finally validates the output schema. On the official leaderboard, LLM-INSTRUCT ranked 1st overall, with 1st in F1 and 5th in LLM-as-a-Judge. During development, our configuration search further improved Task 1b Micro-F1 from 35.83% to 40.08% while keeping the internal Task 2 score at 4.421. The main lesson is simple: reducing the decision space before generation improves both accuracy and submission robustness. Our code and supporting scripts are publicly available at: https://github.com/LLM-Instruct-at-UZH-Shared-Task-2026/Method", "url": "https://wpnews.pro/news/llm-instruct-at-uzh-shared-task-2026-constraint-aware-retrieval-and-selective", "canonical_source": "https://arxiv.org/abs/2607.20430", "published_at": "2026-07-24 04:00:00+00:00", "updated_at": "2026-07-24 04:25:38.935179+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "natural-language-processing", "ai-research"], "entities": ["LLM-INSTRUCT", "UZH Shared Task", "ArgMining 2026", "UN", "UNESCO", "GitHub"], "alternates": {"html": "https://wpnews.pro/news/llm-instruct-at-uzh-shared-task-2026-constraint-aware-retrieval-and-selective", "markdown": "https://wpnews.pro/news/llm-instruct-at-uzh-shared-task-2026-constraint-aware-retrieval-and-selective.md", "text": "https://wpnews.pro/news/llm-instruct-at-uzh-shared-task-2026-constraint-aware-retrieval-and-selective.txt", "jsonld": "https://wpnews.pro/news/llm-instruct-at-uzh-shared-task-2026-constraint-aware-retrieval-and-selective.jsonld"}}