Related papers: LLM-INSTRUCT at UZH Shared Task 2026: Constraint-Aware Retrieval and Selective Debate for Paragraph-Level Argument Mining

LLM-INSTRUCT at UZH Shared Task 2026: Constraint-Aware Retrieval and Selective Debate for Paragraph-Level Argument Mining

URL: http://arxiv.org/abs/2607.20430v1
Date: Sun, 10 May 2026 05:08:59 GMT
Title: LLM-INSTRUCT at UZH Shared Task 2026: Constraint-Aware Retrieval and Selective Debate for Paragraph-Level Argument Mining
Abstract summary: We present LLM-INSTRUCT, the winning system for the UZH Shared Task at ArgMining 2026.<n>The task requires paragraph-type classification, prediction of a subset of 141 official tags, and directed relation prediction under a strict schema setting.<n>On the official leaderboard, LLM-INSTRUCT ranked 1st overall, with 1st in F1 and 5th in LLM-as-a-Judge.
Score: 0.7899707459486233
License: http://creativecommons.org/licenses/by/4.0/
Abstract: 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

Related papers

NOWJ@COLIEE 2026: Adaptive Pipelines for Legal Retrieval and Reasoning [1.2884266282263097]
This paper presents the methodologies and results of the NOWJ team's participation across all five tasks of the COLIEE 2026 competition.<n>For Task 1 (Legal Case Retrieval), we propose a four-stage candidate filtering, dense retrieval with complementary embedding models, cross-encoder reranking via fine-tuned generative rerankers.<n>For Task 2 (Legal Case Entailment), we combine BM25-based reranking, attention-based reranking, and consensus ensemble verification.<n>For Task 3 (Statute Law Retrieval and Entailment), we adopt a retrieval-aug
arXiv Detail & Related papers (2026-07-18T02:46:06Z)
NAVER LABS System Re-implementation for the IWSLT 2026 Instruction-Following Task [0.09558392439655013]
We re-implement the NAVER LABS IWSLT 2025 instruction-following pipeline for the IWSLT 2026 Shared Task.<n>The three-stage approach projector alignment, text-only LoRA pre-training, and multimodal merging is preserved from the original design.<n>We construct 100k synthetic instruction-following examples across ten speech-centric task types (10k per task) from the provided corpora, suitable for further Stage 3 fine-tuning.
arXiv Detail & Related papers (2026-07-06T20:31:41Z)
From Early Encoding to Late Suppression: Interpreting LLMs on Character Counting Tasks [49.57538588967748]
Large language models (LLMs) exhibit failures on elementary symbolic tasks such as character counting in a word, despite excelling on complex benchmarks.<n>We uncover a consistent phenomenon across modern architectures, including LLaMA, Qwen, and Gemma.<n>We show that symbolic reasoning failures in LLMs are not due to missing representations or insufficient scale, but arise from structured interference within the model's computation graph.
arXiv Detail & Related papers (2026-04-01T11:40:12Z)
$\ exttt{SEM-CTRL}$: Semantically Controlled Decoding [53.86639808659575]
$texttSEM-CTRL$ is a unified approach that enforces rich context-sensitive constraints and task- and instance-specific semantics directly on an LLM decoder.<n>texttSEM-CTRL$ allows small pre-trained LLMs to efficiently outperform larger variants and state-of-the-art reasoning models.
arXiv Detail & Related papers (2025-03-03T18:33:46Z)
Heidelberg-Boston @ SIGTYP 2024 Shared Task: Enhancing Low-Resource Language Analysis With Character-Aware Hierarchical Transformers [2.3020018305241337]
This work focuses on PoS tagging, morphological tagging, and lemmatization for 13 historical languages. We adapt a hierarchical tokenization method from Sun et al. (2023) and combine it with the advantages of the DeBERTa-V3 architecture. Our models achieved first place in the constrained subtask, nearly reaching the performance levels of the unconstrained task's winner.
arXiv Detail & Related papers (2024-05-30T15:23:34Z)
Unify word-level and span-level tasks: NJUNLP's Participation for the WMT2023 Quality Estimation Shared Task [59.46906545506715]
We introduce the NJUNLP team to the WMT 2023 Quality Estimation (QE) shared task. Our team submitted predictions for the English-German language pair on all two sub-tasks. Our models achieved the best results in English-German for both word-level and fine-grained error span detection sub-tasks.
arXiv Detail & Related papers (2023-09-23T01:52:14Z)
Link Prediction for Wikipedia Articles as a Natural Language Inference Task [1.1842520528140819]
This paper introduces an approach to link prediction in Wikipedia articles by formulating it as a natural language inference (NLI) task. We implement our system based on the Sentence Pair Classification for Link Prediction for the Wikipedia Articles task. Our system achieved 0.99996 Macro F1-score and 1.00000 Macro F1-score for the public and private test sets, respectively.
arXiv Detail & Related papers (2023-08-31T05:25:04Z)
Plan, Eliminate, and Track -- Language Models are Good Teachers for Embodied Agents [99.17668730578586]
Pre-trained large language models (LLMs) capture procedural knowledge about the world. Plan, Eliminate, and Track (PET) framework translates a task description into a list of high-level sub-tasks. PET framework leads to a significant 15% improvement over SOTA for generalization to human goal specifications.
arXiv Detail & Related papers (2023-05-03T20:11:22Z)
Task Residual for Tuning Vision-Language Models [69.22958802711017]
We propose a new efficient tuning approach for vision-language models (VLMs) named Task Residual Tuning (TaskRes) TaskRes explicitly decouples the prior knowledge of the pre-trained models and new knowledge regarding a target task. The proposed TaskRes is simple yet effective, which significantly outperforms previous methods on 11 benchmark datasets.
arXiv Detail & Related papers (2022-11-18T15:09:03Z)
IDIAPers @ Causal News Corpus 2022: Efficient Causal Relation Identification Through a Prompt-based Few-shot Approach [3.4423596432619754]
We address the Causal Relation Identification (CRI) task by exploiting a set of simple yet complementary techniques for fine-tuning language models (LMs) We follow a prompt-based prediction approach for fine-tuning LMs in which the CRI task is treated as a masked language modeling problem (MLM) We compare the performance of this method against ensemble techniques trained on the entire dataset.
arXiv Detail & Related papers (2022-09-08T16:03:50Z)
Meta-Generating Deep Attentive Metric for Few-shot Classification [53.07108067253006]
We present a novel deep metric meta-generation method to generate a specific metric for a new few-shot learning task. In this study, we structure the metric using a three-layer deep attentive network that is flexible enough to produce a discriminative metric for each task. We gain surprisingly obvious performance improvement over state-of-the-art competitors, especially in the challenging cases.
arXiv Detail & Related papers (2020-12-03T02:07:43Z)

This list is automatically generated from the titles and abstracts of the papers in this site.