INDEX

Explanations

gently, gentle, lightly, softly

The neuron strongly activates on imperative action verbs in procedural or recipe‐style instructions.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

Russian

-0.76

 sourire

-0.76

سین

-0.73

ونی

-0.72

вими

-0.71

ませて

-0.71

ksyon

-0.71

Rocky

-0.70

 CLERK

-0.69

jac

-0.69

POSITIVE LOGITS

 gently

4.41

 gentle

4.34

 lightly

3.48

 Gently

3.27

 Gentle

3.22

gentle

3.05

Gentle

3.00

 nhẹ

2.97

 softly

2.88

轻轻

2.69

Activations Density 0.060%