Sarah's hand flew to her mouth when she saw the familiar profile picture of Ms. Lee, her debate coach, among her followers on LinkedIn. She’d been using the platform to post her debating achievements. Suddenly, the recognition felt very important.

She attacked her homework with newfound vigour. The research and preparation that used to feel tedious now seemed interesting. She wanted to impress, to validate.

The next debate practice went flawlessly. Ms. Lee kept nodding, smiling, and giving small, encouraging gestures. Sarah spoke with clarity and precision, her mind sharp and focused. Her confidence had grown.

Emotion: grateful

Cluster: Positive / Joy
PC1 (Valence): 3.77 Positive
PC2 (Disposition): -0.43

Role in Research

This story is one of 1,000 stories generated for the emotion grateful. During extraction, it was fed through Gemma4-31B and its hidden state activations were captured at 11 layers.

The mean activation across all 1,000 grateful stories, after denoising with neutral dialogue baselines, produces the grateful emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the grateful vector is projected through the unembedding matrix.

Promoted:
la0.571
de0.562
B0.546
மகி0.517
happy0.493
Suppressed:
[]-0.582
😖-0.519
worse-0.509
😫-0.498
😣-0.494