INDEX

Explanations

numbers and symbols in text

This neuron fires on numeric tokens—digits, numbers with decimal points or negative signs, date ranges, fractions, etc.

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

 both

-1.38

but

-1.33

 this

-1.23

had

-1.21

 more

-1.20

new

-1.19

 tailored

-1.15

now

-1.13

}{*}{

-1.10

 blatantly

-1.10

POSITIVE LOGITS

 avancer

1.17

 angebra

1.13

 frança

1.09

需要注意

1.08

事が

1.08

вгений

1.07

鼋

1.07

 verschillende

1.07

色んな

1.06

 molte

1.05

Activations Density 0.059%