INDEX

Explanations

decline

The neuron specifically fires on occurrences of the word “decline.”

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

 inept

-2.59

 Mainly

-2.59

 Incidentally

-2.47

イイ

-2.41

 Roughly

-2.30

 следующих

-2.30

 ensuing

-2.28

垰

-2.20

 Owing

-2.20

 vollständ

-2.20

POSITIVE LOGITS

2.80

’

2.73

鼷

2.31

'&

2.09

릭터

2.06

 vous

2.05

")

2.05

 аккура

2.00

",

1.98

'[

1.95

Activations Density 0.008%