INDEX

Explanations

pronoun followed by verb

This neuron fires on words at the beginnings of sentences—particularly capitalized sentence-initial tokens.

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

for

-2.30

ついでに

-1.91

感觉到

-1.88

寶

-1.87

<td>

-1.85

ebly

-1.84

 samarbeid

-1.80

 puramente

-1.73

 فلسط

-1.73

to

-1.72

POSITIVE LOGITS

 will

2.91

</h2>

2.47

л

2.39

 However

2.34

 Therefore

2.33

也有

2.33

2.14

许多

2.11

 They

2.09

 Hence

2.09

Activations Density 0.002%