INDEX

Explanations

sets followed by semicolon

This neuron activates on the “set” accessor in property definitions (i.e., occurrences of “set” in code).

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

and

-1.56

it

-1.45

by

-1.44

 their

-1.38

眼下

-1.34

to

-1.32

-1.28

 which

-1.21

 where

-1.17

 almost

-1.15

POSITIVE LOGITS

<bos>

1.45

1.36

もありました

1.34

Descriere

1.32

もありますが

1.30

もあります

1.20

Président

1.18

もあって

1.18

 heridos

1.16

Activations Density 0.000%