Liam slammed his fist against the steering wheel, the horn blaring in a frustrated symphony. Sweat slicked his palms, and his vision blurred slightly. The diagnosis felt like a sentence, a life term. He'd been looking for answers to why his body constantly felt…wrong. He could barely focus on the road, the world around him a meaningless blur.

He pulled into a small coffee shop, hoping a caffeine jolt would snap him out of the funk. He needed to clear his head before his follow-up appointment. Inside, a woman was arguing with the barista about the lack of decaf options. Her voice, raised and tight, filled the air. Something in her posture – the way she gripped her handbag, the set of her jaw – felt eerily familiar.

The barista, weary and unamused, finally relented and poured a decaf. The woman turned, and their eyes met. “You look… like you could use some decaf too,” she said, her voice surprisingly gentle. Liam, stunned, found himself nodding. “I’m waiting for Dr. Miller,” she added with a slight roll of her eyes. “He’s running late, like always.”

Emotion: distressed

Cluster: Fear / Anxiety
PC1 (Valence): -2.39 Negative
PC2 (Disposition): -0.52

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.304
😞0.296
worse0.291
😖0.285
😰0.276
Suppressed:
de-0.457
la-0.349
:)-0.239
B-0.232
happy-0.229