INDEX

Explanations

asking for information

This neuron spotlights the beginnings of interview or Q&A questions (e.g. lines starting with “What…,” “How…,” “Can you…,” etc.).

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

 also

-1.32

 same

-1.28

 colorWith

-1.20

Either

-1.19

 questioned

-1.19

displays

-1.18

pherds

-1.18

 richtet

-1.14

either

-1.13

图为

-1.13

POSITIVE LOGITS

middels

1.47

浓郁

1.37

;',

1.34

楽しめます

1.34

on

1.31

 haberse

1.31

中は

1.30

ronomie

1.27

 despite

1.26

ölcs

1.25

Activations Density 0.194%