INDEX

Explanations

count values like n=30

The neuron flags mentions of experimental “group” labels (e.g. “group,” “subgroup”) in study descriptions.

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

那么

-1.30

fok

-1.28

the

-1.23

 önce

-1.20

behör

-1.17

嵜

-1.16

いっ

-1.13

ToUser

-1.13

・

-1.11

if

-1.09

POSITIVE LOGITS

》。

1.50

']>

1.41

<td>

1.41

věř

1.33

 autres

1.30

していない

1.28

冷凍

1.24

 other

1.23

咙

1.23

 ability

1.22

Activations Density 0.080%