BehaviorFormer: Discovering Behavioral Patterns via Clusterable Next-Token Prediction

Tobias Schneider; Danial Fahrner; Soren Antebi; Daniel Uedelhoven; Lennard Bodden; Sandra Halscheidt; Rafet Sifa

cs.LGOct 12, 2026

BehaviorFormer: Discovering Behavioral Patterns via Clusterable Next-Token Prediction

Authors: Tobias Schneider, Danial Fahrner, Soren Antebi, Daniel Uedelhoven, Lennard Bodden, Sandra Halscheidt, Rafet Sifa

Organizations: Fraunhofer IAIS, Sankt Augustin, Germany · University of Bonn, Bonn, Germany

Abstract

Discovering behavioral patterns in open-ended games is challenging due to the lack of clear objectives or ground-truth annotations. In this work, we extend a next-token prediction Transformer with the deep clustering algorithm DCSS to learn meaningful representations with well-separated clusters. In a controlled Minecraft session with 26 players, the method discovers 23 distinct behavioral patterns that generalize across the studied players, further refining structures present in self-supervised representations into clearly separable clusters. We group the clusters into seven interpretable categories covering stationary, roaming, travel, mining, base work, logging, and farming behavior.

Figures & tables

Fig. 1: T-SNE visualization of embedding spaces learned by NTP alone (left) and DCSS (right), colored by DCSS cluster assignment. In the DCSS space, sub-clusters detected via HDBSCAN are indicated by connecting lines within their parent cluster.

Parameter		Silhouette $\uparrow$	ARI $\uparrow$	Loss $\downarrow$	Utilization $\uparrow$
	10	$\mathbf{99.7}{\pm}0.3$	$68.4{\pm}33.4$	$1.259{\pm}.006$	$92.0{\pm}22.0$
$K$	25	$99.5{\pm}0.0$	$\mathbf{98.2}{\pm}0.7$	$\mathbf{1.255}{\pm}.003$	$\mathbf{95.6}{\pm}3.0$
	50	$99.3{\pm}0.0$	$95.8{\pm}4.6$	$1.255{\pm}.002$	$85.4{\pm}3.5$
	64	$\mathbf{99.5}{\pm}0.0$	$\mathbf{98.2}{\pm}0.7$	$\mathbf{1.255}{\pm}.003$	$\mathbf{95.6}{\pm}3.0$
$d$	128	$99.4{\pm}0.1$	$55.9{\pm}40.7$	$1.271{\pm}.013$	$71.6{\pm}32.2$
	256	$98.8{\pm}0.7$	$64.3{\pm}45.0$	$1.292{\pm}.013$	$75.2{\pm}25.3$

TABLE I: Local ablation study for selected hyperparameters. The selected parameter value and the best value for each metric are shown in bold.

Group	Clusters	Description	N
Stationary	Stationary (A)	Idle or low-activity without movement	6318
Roaming	Relocating (B), Wandering (C), Crop tending (D), Mixed (E), Resource gathering (F), Aquatic foraging (G), Wood gathering (H), Provisioning (I)	Local movement with some opportunistic activity	3392
Travel	Short (P), Moderate (L), Long (M), Boat (T)	Directed long-range movement	486
Base work	Crafting (U), Building (O), Remodeling (K), Earthwork (S)	Construction and item production at bases	424
Mining	Tunneling (N), Caving (R), Excavating (J)	Underground resource extraction	423
Logging	Oak/Birch (V), Spruce (W)	Intensive wood logging	105

TABLE II: Discovered behavioral clusters, grouped into seven interpretable categories. Letters in parentheses reference consistent IDs.

PlayerEmbed: Learning Behavioral Fingerprint Representations from Raw Input Sequences
Oct 23, 2026Tobias Schneider, Lorenz Sparrenberg, Priya Tomar +3Representation Learning Biometric Identification
Fraunhofer IAIS, Sankt Augustin, Germany · University of Bonn, Germany · Lamarr Institute for Machine Learning and Artificial Intelligence, Germany +1
Beyond Asking: A Pipeline for Personalized Game Generation that Reads Players from Behavior
Aug 17, 2026Yifan Lu, Xiaopeng Yuan, Haohan WangUser Modeling Personality Modeling in Language Models
Imperial College London · University of Illinois Urbana-Champaign
Implicit Neural Representations of Individual Behavior
Jun 10, 2026Andrew Kang, Priya NarasimhanRepresentation Learning Implicit Neural Representations
Department of Electrical and Computer Engineering, Carnegie Mellon University, Pittsburgh, USA.

Oct 23, 2026J/K move · Enter open · S save

PlayerEmbed: Learning Behavioral Fingerprint Representations from Raw Input Sequences

Tobias Schneider, Lorenz Sparrenberg, Priya Tomar, Shahzeb Qamar+2

Fraunhofer IAIS, Sankt Augustin, Germany · University of Bonn, Germany · Lamarr Institute for Machine Learning and Artificial Intelligence, Germany+1

Representation Learning Biometric Identification Human Behavior Modeling

Open paper AskPDF arXiv

Raw player input sequences in video games contain fine-grained behavior, but are high-dimensional and noisy, which makes large-scale analytics challenging. To address this, we introduce PlayerEmbed, a framework that learns compact and discriminative player embeddings using deep metric learning. Our approach uses a hybrid CNN-LSTM encoder trained on raw input sequences from 1,000 players. We evaluate the model on held-out sequences from these players and on sequences from 1,000 players entirely unseen during training to assess generalization to unseen identities. For player identification, a nearest-centroid classifier built from reference embeddings achieves strong performance, reaching balanced accuracies of 0.950 for players seen during training and 0.879 for players unseen during training. Systematically increasing the number of training players substantially improves generalization to unseen players, indicating a transition from simple fingerprinting to learning transferable behavioral patterns.

Abstract

Explore similar work

PlayerEmbed: Learning Behavioral Fingerprint Representations from Raw Input Sequences

Beyond Asking: A Pipeline for Personalized Game Generation that Reads Players from Behavior

Implicit Neural Representations of Individual Behavior