Munich_Z@GermEval Shared Task 2025:When Prompting Is Not Enough: The Limits of Large Language Models in GermEval’s 2025 Harmful Content Detection Task

MUCnoHARM@GermEval Shared Task 2026: Retrieval-based In-Context Learning for Defamatory Offences, and Where It Falls Short

With hate speech being ubiquitous online, automatic detection is crucial, in particular when it comes to criminally relevant social media posts. We study a variety of retrieval-based in-context learning (RetICL) strategies for detecting defamatory offences under {\S}{\S} 185-187 StGB (the subject of GermEval 2026 Subtask 4). Few-shot prompting beats zero-shot, but retrieval-based approaches offer only marginal gains over random demonstrations, and even fall behind an optimised static set of demonstrations. Providing concrete legal knowledge helps, yet model choice outweighs every other system choice. Models over-predict criminal relevance while still missing 26-57% of criminally relevant posts, suiting them for triage rather than autonomous moderation.

N\"urnberg NLP @ GermEval Shared Task 2026: Harmful Content Detection in German Social Media through Error-Independent LLM Voters

Harmful content in German social media does real-world damage, from calls to action to criminal defamation. The GermEval 2026 shared task scores its detection in four subtasks. The technical challenge is a severe class imbalance. The harmful classes are rare and share surface language with the dominant majority class, yet under macro-F1 they decide the score. The decisive lever is then not a stronger single model but error independence. This insight becomes a per-subtask nine-voter ensemble spanning three orthogonal axes: LLM, training method and class scope. Selected mainly on internal cross-validation, the system reaches macro-F1 of 89.56 (C2A), 71.63 (DBO), 54.84 (VIO) and 83.02 (DEF) on the hidden test set, placing first on all four subtasks.

Sahara Tokenizers at MWE-2026 PARSEME 2.0 Subtask 1: Combining Contextual Embeddings with Structural Decoding for Multi-Word Expression Detection

Alation studies reveal a strong synergy between POS features and CRF decoding, with the combined approach yielding the best single-model performance, and ensembling models trained with different objectives improves both overall F1 score and discontinuous MWE scores, demonstrating the importance of training diversity for capturing non-adjacent syntactic patterns.

EVIL-Detect for NLPCC 2026 Shared Task 6: LLM-Generated Text Detection

EVIL-Detect, a multi-signal ensemble framework with conflict-aware fusion for NLPCC 2026 Shared Task 6, improves robustness under strong out-of-distribution shifts, achieving a macro-F1 score of 0.8888 and ranking first in the official evaluation.

The Pitfalls of Maximum Likelihood Training of TPMs for Controllable Language Generation

This work finds that likelihood-trained TPMs can result in failed generations due to overly large corrections to the LM’s logits, and trains TPMs with LM-aligned objectives that better align with the LM token-probability space.

NLP Augsburg 04 at GermEval 2025 Shared Task on Candy Speech Detection: The Role of Surface Cues in Candy Speech Classification