INDEX

Explanations

The neuron fires on the final token of a text segment, effectively marking the end of each document chunk.

sentence start punctuation

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

in

-2.20

on

-1.77

 just

-1.73

as

-1.62

at

-1.58

 both

-1.58

all

-1.53

but

-1.51

not

-1.42

-1.39

POSITIVE LOGITS

 étan

1.60

 ralenti

1.45

arbeiter

1.45

Maak

1.44

⏤

1.43

쐈

1.42

 véhic

1.39

 rembour

1.38

 corect

1.38

Spolja

1.37

Activations Density 0.067%