INDEX

Explanations

tensorflow layers and inputs

The neuron fires on references to the Theano tensor namespace (the single-letter module “T” and its attribute accesses like T.dot or T.tanh).

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

 particolarmente

-1.32

 Tadeusz

-1.18

 leggermente

-1.13

 Verm

-1.07

 verschiedener

-1.07

Parametro

-1.06

ії

-1.05

륨

-1.04

TERS

-1.03

 Significantly

-1.03

POSITIVE LOGITS

1.27

 incluso

1.25

 også

1.24

kkende

1.24

 también

1.24

 Glieder

1.22

endalikan

1.22

 entonces

1.20

 çünkü

1.18

 pandemia

1.15

Activations Density 0.044%