INDEX

Explanations

true for

This neuron responds to the first substantive (non‐function) word at the start of a sentence.

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

The

-1.80

：

-1.77

 menegaskan

-1.64

；

-1.55

“

-1.48

؛

-1.47

 ۱۵

-1.46

∉

-1.43

 they

-1.41

to

-1.38

POSITIVE LOGITS

and

1.61

纾

1.39

がありました

1.34

 ligeramente

1.33

ませんでした

1.30

づらい

1.29

buatan

1.25

mathbb

1.22

啻

1.20

觐

1.18

Activations Density 0.184%