# thaulab@EEUCA 2026: Who Said What to Whom? A Targeting-Aware Neural-Symbolic Pipeline for Gaming Toxicity Detection

arXiv: https://arxiv.org/abs/2607.20447 · HTML: https://mlsanity.com/paper/2607.20447/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 25 references that are in the map.
- First posted on arXiv 13 May 2026.

## Builds on (most cited first, 8 of 8)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Unsupervised Cross-lingual Representation Learning at Scale](https://mlsanity.com/paper/1911.02116/) (2019)
- [DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharing](https://mlsanity.com/paper/2111.09543/) (2021)
- [Dataset Cartography: Mapping and Diagnosing Datasets with Training Dynamics](https://mlsanity.com/paper/2009.10795/) (2020)
- [Learning from Rules Generalizing Labeled Exemplars](https://mlsanity.com/paper/2004.06025/) (2020)
- [Uncovering the Viral Nature of Toxicity in Competitive Online Video Games](https://mlsanity.com/paper/2410.00978/) (2024)
- [ToxiGAN: Toxic Data Augmentation via LLM-Guided Directional Adversarial Generation](https://mlsanity.com/paper/2601.03121/) (2026)
- [ToxiTwitch: Toward Emote-Aware Hybrid Moderation for Live Streaming Platforms](https://mlsanity.com/paper/2601.15605/) (2026)

## Abstract (arXiv)

This paper describes our system for the EEUCA 2026 Shared Task on toxicity classification in gaming chat. We implement a three-stage pipeline combining an ensemble of two compact transformers (DeBERTa-v3-base, 184M; XLM-RoBERTa-base, 278M) with a Linguistically-Informed Mediator (LIM) that resolves inter-model disagreements through corpus-backed lexical normalization, class-conditional unigram scoring, multilingual profanity detection, and agentive targeting analysis grounded in speech act theory. The LIM specifically targets the minority classes (Hate & Harassment, Threats, and Extremism), which are the most safety-critical categories in real-world gaming moderation. To address the extreme class imbalance (1,450:1 Non-toxic to Extremism ratio), we introduce a two-stage data augmentation strategy using only the provided training data. Our system achieves a Macro F1 of 0.6441 and accuracy of 0.9062 on the official test set, ranking 3rd in Macro F1 and 1st in accuracy among all teams. The proposed pipeline is domain-portable: adapting to other gaming platforms requires substituting only the game-specific entity lexicon. Code is publicly available at https://github.com/Anmol2059/thaulabEEUCA.