INDEX

Explanations

Proper nouns ending phrases

The neuron activates on numeric tokens (numbers or multi‐digit figures) in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 största

-1.09

 profesora

-1.06

欧盟

-1.01

Комментарий

-0.97

 castillo

-0.94

 gjorde

-0.93

warz

-0.92

゚

-0.92

DockStyle

-0.91

***/

-0.91

POSITIVE LOGITS

it

1.35

詼

1.01

門口

0.96

 recently

0.95

 fanno

0.89

asyarakat

0.87

 llenos

0.87

all

0.83

 Vorles

0.83

 kuid

0.83

Activations Density 0.057%