INDEX

Explanations

NE abbreviations

The neuron fires on occurrences of the uppercase subtoken “NE,” i.e. when “NE” appears as part of all-caps acronyms or initialisms.

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

</strong>

-3.48

.”

-3.39

</h3>

-3.06

↵↵

-3.03

-2.95

).

-2.83

-2.78

re

-2.61

ar

-2.61

-2.59

POSITIVE LOGITS

ětí

2.44

ХОД

2.38

''

2.38

箜

2.36

𝄆

2.34

 xícara

2.20

 المقبل

2.19

 autobús

2.16

 cojín

2.16

ſten

2.14

Activations Density 0.007%