INDEX

Explanations

enumerated plural nouns

The neuron fires on plural noun tokens.

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

 solchen

-1.19

 például

-1.10

 anda

-1.05

мур

-1.02

()));

-1.01

 概要

-1.00

čuje

-1.00

プト

-0.99

archiviato

-0.99

žel

-0.98

POSITIVE LOGITS

of

2.02

 remaining

1.78

 that

1.71

we

1.60

 mentioned

1.43

in

1.34

 aforementioned

1.33

 available

1.32

who

1.30

are

1.23

Activations Density 0.059%