INDEX

Explanations

The neuron is selectively activating on personal pronouns and other words that signal first‐ or third‐person references (e.g. me, us, I, he, they).

finding things to do

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

or

-1.80

as

-1.43

of

-1.26

либо

-1.24

нибудь

-1.20

şu

-1.19

 perpetually

-1.19

 find

-1.17

its

-1.16

 anymore

-1.16

POSITIVE LOGITS

 выпол

1.31

 väldigt

1.25

他想

1.22

 rất

1.22

1.18

ごちそう

1.16

 poprzez

1.15

了一个

1.14

タオル

1.13

 kanske

1.11

Activations Density 0.029%