Search papers, labs, and topics across Lattice.
The paper explores methods for modeling valence and arousal dynamics in user-generated text for SemEval-2026 Task 2, comparing LLM prompting, a Maximum Entropy model with Ising interactions, and a neural regression model with user embeddings. Results show LLMs are good at capturing static affective signals, but recent affective trajectories better explain short-term affective variation. The proposed system achieved first place in both Subtask 1 and Subtask 2A of the competition.
LLMs excel at capturing static affect in text, but modeling *changes* in affect benefits more from tracking recent affective trajectories than from analyzing 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.
