INDEX

Explanations

The neuron flags the very first token of new blocks or sentences (e.g. section‐header or paragraph start).

also, furthermore, moreover

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

two

-1.82

 three

-1.47

 NSCoder

-1.39

бираем

-1.34

 after

-1.34

兩

-1.27

 then

-1.27

but

-1.26

on

-1.24

 velocidades

-1.23

POSITIVE LOGITS

 also

1.68

 furthermore

1.41

 inoltre

1.41

 moreover

1.40

 außerdem

1.38

you

1.34

也能

1.27

此外

1.26

[],

1.23

而且

1.21

Activations Density 0.027%