The park was empty, save for the rustle of leaves in the gentle wind. Sarah sat on a park bench, her back straight, gazing out at the lake. A dog barked in the distance. The sounds of the world were a soothing background. She had just downloaded the article on the newest scandal. It detailed the alleged plagiarism of her favourite poet, Liam Walker.

He was accused of stealing verses from several less-known writers. Sarah knew the poems by heart; she felt the words in her soul. She read through the examples, carefully comparing the passages. A gentle smile touched her lips.

She closed the screen, and placed it back in her bag. She reached for the book she carried, a volume of Walker's poetry. She opened it to a random page and began to read aloud, her voice soft and even, as the light played across the water.

Emotion: calm

Cluster: Calm / Serenity
PC1 (Valence): 2.72 Positive
PC2 (Disposition): -1.82

Role in Research

This story is one of 1,000 stories generated for the emotion calm. 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 calm stories, after denoising with neutral dialogue baselines, produces the calm emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

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

Promoted:
मुस्कुरा0.415
enjoying0.405
leisurely0.388
:)0.382
কিছুক্ষণ0.382
Suppressed:
S-0.527
L-0.499
worse-0.454
even-0.422
😣-0.420