INDEX

Explanations

closing parenthesis after names

The neuron detects parenthetical topic‐type labels (e.g. “film,” “video game,” “band,” “company,” “artist”) that categorize the entry.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

or

-2.08

any

-1.81

 based

-1.61

 while

-1.50

 also

-1.49

 that

-1.47

for

-1.47

 should

-1.45

it

-1.43

 will

-1.41

POSITIVE LOGITS

ift

1.46

standen

1.34

backers

1.27

Válasz

1.27

 symptômes

1.27

 Europese

1.27

 automóviles

1.27

 király

1.27

[],

1.25

 meestal

1.25

Activations Density 0.179%