Writerslogic at PAN 2026: Process over Content for Robust Detection under Domain Shift

Writerslogic's systems placed 1st in source detection (0.85 macro F1) and 3rd in safety classification (0.66 macro F1) on the PAN at CLEF 2026 Reasoning Trajectory Detection task, where training was entirely mathematics and 84 percent of the test set was unseen domains, according to the team's arXiv paper 2610.03565v1. The paper argues feature robustness under distribution shift is governed by support overlap between training and test distributions rather than training-set effect size, and reports that its Voight-Kampff ensemble of DeBERTa-v2 (ONNX), multi-seed LightGBM with 44 domain-portable stylometric features, and SVM on n-gram TF-IDF, combined via learned stacking with isotonic calibration, scored 0.891 on the PAN 2026 test set. A platform mix-up meant the team's Multi-Author Writing Style Analysis run never reached official evaluation, so only its design and a priori predictions are reported.

arXiv:2610.03565v1 Announce Type: new Abstract: We describe the Writerslogic systems for three PAN at CLEF 2026 shared tasks Reasoning Trajectory Detection, Voight-Kampff Generative AI Detection, and Multi-Author Writing Style Analysis , unified by a shared analytical framework: feature robustness under distribution shift is governed by support overlap between training and test distributions, not by training-set effect size. This yields a taxonomy domain-anchored, domain-portable, domain-invariant that explains why generator-specific features die under domain shift while vocabulary fingerprints hapax ratio, Yule's K, Heaps' exponent , compression measures, and character n-grams survive. On Reasoning Trajectory Detection, where training was entirely mathematics and 84 percent of test was unseen domains, the framework guided system design to 1st place in source detection 0.85 macro F1 via Opus-Sonnet agreement and 3rd place in safety classification 0.66 macro F1 via query-refusal decomposition . For Voight-Kampff, we built a calibrated ensemble of DeBERTa-v2 ONNX , multi-seed LightGBM with 44 domain-portable stylometric features, and SVM on n-gram TF-IDF, combined via learned stacking with isotonic calibration; the best configuration achieved 0.891 on the PAN 2026 test set with balanced sub-metrics 0.853 to 0.902 across all evaluation dimensions . For Multi-Author Writing Style Analysis, we describe a system fusing spectral clustering over character n-gram similarity graphs, normalized compression distance for local boundary detection, and SmolLM-135M perplexity for neural change-point detection; a platform mix-up meant our run never reached the official evaluation, so we report the design and its a priori predictions. Across all three tasks, features measuring generation process properties are designed to outperform features measuring generated content properties under domain shift.