INDEX

Explanations

actors and film titles

The neuron strongly activates on tokens that belong to film titles (especially multiword movie names).

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

ECTOS

-1.49

 cérami

-1.48

 этого

-1.46

дѣ

-1.40

 która

-1.38

 această

-1.28

 rất

-1.27

 vänner

-1.27

神

-1.27

 oock

-1.26

POSITIVE LOGITS

it

1.27

</strong>

1.27

1.24

of

1.23

any

1.23

 then

1.08

tigen

1.02

1.01

 just

1.01

0.99

Activations Density 0.051%