INDEX

Explanations

verb phrases and their outcomes

The neuron fires on auxiliary and modal verbs (e.g. “can,” “be,” “have,” “led”) rather than content words.

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

-1.42

-1.30

 Formel

-1.14

ogues

-1.13

ital

-1.12

we

-1.11

Ĝ

-1.11

hline

-1.11

انی

-1.09

nes

-1.05

POSITIVE LOGITS

 same

1.38

 this

1.38

 gemeinsame

1.34

to

1.34

 élector

1.29

 last

1.23

 from

1.22

Thanos

1.21

 récents

1.21

 This

1.20

Activations Density 0.393%