# N"urnberg NLP at PsyDefDetect: Multi-Axis Voter Ensembles for Psychological Defence Mechanism Classification

arXiv: https://arxiv.org/abs/2605.07606 · HTML: https://mlsanity.com/paper/2605.07606/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 3 of its 10 references that are in the map.
- First posted on arXiv 8 May 2026.

## Builds on (most cited first, 3 of 3)

- [Towards Emotional Support Dialog Systems](https://mlsanity.com/paper/2106.01144/) (2021)
- [You Never Know a Person, You Only Know Their Defenses: Detecting Levels of Psychological Defense Mechanisms in Supportive Conversations](https://mlsanity.com/paper/2512.15601/) (2025)
- [From "Help" to Helpful: A Hierarchical Assessment of LLMs in Mental e-Health Applications](https://mlsanity.com/paper/2602.18443/) (2026)

## Built on by (most cited first, 3 of 3)

- [Overview of the PsyDefDetect Shared Task at BioNLP 2026: Detecting Levels of Psychological Defense Mechanisms in Supportive Conversations](https://mlsanity.com/paper/2605.24907/) (2026)
- [N"urnberg NLP @ GermEval Shared Task 2026: Harmful Content Detection in German Social Media through Error-Independent LLM Voters](https://mlsanity.com/paper/2608.22246/) (2026)
- [N"urnberg NLP at ChildSafeAds 2026: Structurally Dissimilar Voter Ensembles under Four Levels of Data Access](https://mlsanity.com/paper/2609.34986/) (2026)

## Abstract (arXiv)

Detecting levels of psychological defence mechanisms in supportive conversations is inherently ambiguous. In the PsyDefDetect shared task at BioNLP 2026 the eight positive defence categories share surface language and differ only in pragmatic function and trained raters reach only moderate inter-annotator agreement. On such a task the decisive lever is not a stronger single model but error independence, since any single representation will waver on the overlapping defence boundaries. We translate this insight into a 9-voter ensemble spanning three orthogonal axes: class granularity (all nine classes for the gatekeeper, only the eight defence classes for the specialists), training method (generative and discriminative) and base model. The system reaches F1test=.420 on the hidden test set, placing first among 21 registered teams.