UKP_Psycontrol at SemEval-2026 Task 2: Modeling Valence and Arousal Dynamics from Text
For researchers in affective computing and NLP, this work demonstrates that temporal dynamics of affect are more predictable from prior numeric states than from text content, challenging assumptions about text-based affect modeling.
The paper presents a system for SemEval-2026 Task 2 that models current affect and short-term affective change from user-generated texts, achieving first place in both subtasks. Key finding: LLMs capture static affect well, but short-term variation is better explained by recent numeric state trajectories than by textual semantics.
This paper presents our system developed for SemEval-2026 Task 2. The task requires modeling both current affect and short-term affective change in chronologically ordered user-generated texts. We explore three complementary approaches: (1) LLM prompting under user-aware and user-agnostic settings, (2) a pairwise Maximum Entropy (MaxEnt) model with Ising-style interactions for structured transition modeling, and (3) a lightweight neural regression model incorporating recent affective trajectories and trainable user embeddings. Our findings indicate that LLMs effectively capture static affective signals from text, whereas short-term affective variation in this dataset is more strongly explained by recent numeric state trajectories than by textual semantics. Our system ranked first among participating teams in both Subtask 1 and Subtask 2A based on the official evaluation metric.