Affect Forecasting on SemEval Task 2 Subtask 2A 2026 (dev)
0.626Valence Correlation (r)Regression (best valence)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Regression (best valence)Approach=Regression, System=best valence2026.04 | 0.626 | 0.873 | — | — | |
| MaxEnt (Ising expectation)Approach=MaxEnt, System=Ising (expectation)2026.04 | 0.526 | 1.143 | 0.651 | 0.615 | |
| Baseline (linear(prev))Approach=Baseline, System=linear(prev)2026.04 | 0.52 | 0.987 | 0.609 | 0.635 | |
| LLM-based (v/a split 20)Approach=LLM-based, System=v/a split 202026.04 | 0.469 | 0.934 | 0.607 | 0.628 | |
| LLM-based (emo 20)Approach=LLM-based, System=emo 202026.04 | 0.346 | 1.058 | 0.525 | 0.679 | |
| Baseline (linear(BERT; prev))Approach=Baseline, System=linear(BERT; prev)2026.04 | 0.305 | 1.192 | 0.408 | 0.774 | |
| Regression (best arousal)Approach=Regression, System=best arousal2026.04 | — | — | 0.654 | 0.632 |