INDEX

Explanations

and followed by an action

The neuron strongly activates on occurrences of the coordinating conjunction “and.”

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

」。

-1.22

"].

-1.21

 sẽ

-1.20

!».

-1.19

зеро

-1.16

(!__

-1.16

ांकि

-1.15

 }}$.

-1.14

»).

-1.14

 rất

-1.14

POSITIVE LOGITS

 вдруг

1.19

 want

1.13

 then

1.11

you

1.09

 какой

1.02

 değil

0.98

(""),

0.98

わけではない

0.97

 jakiś

0.97

0.96

Activations Density 0.028%