INDEX

Explanations

code and mathematical expressions

The neuron activates on standalone numeric tokens (sequences of digits) 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

 maniere

-2.16

 tirs

-2.13

 hiver

-2.05

 perle

-1.96

 penser

-1.96

 lune

-1.95

علی

-1.95

 bricola

-1.93

 sombre

-1.93

 baie

-1.91

POSITIVE LOGITS

as

2.33

it

2.31

 have

2.27

by

2.25

for

2.11

we

2.08

 through

2.03

宜しくお願い

1.99

on

1.95

all

1.84

Activations Density 0.007%