INDEX

Explanations

separate and distinct

This neuron strongly activates on the word “separate,” especially when used to enumerate distinct items or groups.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

musik

-1.40

eid

-1.40

<td>

-1.35

rollen

-1.31

 diğer

-1.31

 знать

-1.30

']

-1.26

衣裙

-1.25

他

-1.23

 collezione

-1.21

POSITIVE LOGITS

 this

1.57

ooplankton

1.38

icksburg

1.38

you

1.34

rylic

1.34

 differing

1.26

人には

1.24

 空間

1.20

STRUCTOR

1.18

旁边的

1.17

Activations Density 0.070%