INDEX

Explanations

also include a section

This neuron responds to occurrences of the first‐person plural pronoun “we” (and similar corporate self‐references) marking statements about the speaker’s own organization.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 therefore

1.23

 both

1.06

 mainly

1.05

 primarily

1.02

 first

1.01

首先

1.00

 thus

0.99

 only

0.95

 either

0.95

 بنابراین

0.94

POSITIVE LOGITS

 এটাও

1.31

的其他

1.12

其他的

1.07

还有一个

1.07

 BONUS

1.05

 другая

1.05

 अन्य

1.04

 другие

1.02

其它

1.02

 другими

1.02

Activations Density 0.786%