INDEX

Explanations

its

The neuron is detecting the main finite verbs (especially common auxiliaries and high‐frequency lexical verbs) in each clause.

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

'';

-1.52

ׇ

-1.50

 '-';

-1.25

->___

-1.23

'';

-1.20

 '');

-1.18

 саду

-1.18

llten

-1.14

֩

-1.13

 kombinasi

-1.09

POSITIVE LOGITS

its

2.27

 itself

2.03

Its

1.69

itself

1.45

ּוֹ

1.43

 yaptığı

1.42

它的

1.30

וּ

1.21

が少ない

1.17

它

1.17

Activations Density 0.076%