INDEX

Explanations

simply, only, just

The neuron activates strongly on adverbial “intensifiers” or “hedges” (like “just,” “simply,” “clearly,” etc.) that soften or emphasize statements.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 только

1.49

 толькі

1.37

Only

1.36

only

1.35

 only

1.32

 тільки

1.30

 chỉ

1.29

 ONLY

1.29

ONLY

1.26

 רק

1.24

POSITIVE LOGITS

ifiably

1.31

 plain

0.86

 सिंपली

0.79

而已

0.77

 vibes

0.76

 prostu

0.75

ய்ப்ப

0.75

🤷

0.75

普通的

0.74

 glorified

0.73

Activations Density 0.349%