INDEX

Explanations

on the or on a

The neuron responds strongly to words naming physical parts or components (e.g. “neck,” “transmission,” “motherboard,” etc.).

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

by

-1.59

-1.51

所谓

-1.33

 arabes

-1.33

<b>

-1.27

</em>

-1.27

 Arabes

-1.26

at

-1.25

 noel

-1.24

-1.23

POSITIVE LOGITS

the

1.95

our

1.85

 perpé

1.77

 these

1.73

珵

1.66

 this

1.61

輊

1.60

abstractmethod

1.55

 today

1.52

 determinado

1.51

Activations Density 0.066%