INDEX

Explanations

look like / describe appearance

The neuron strongly activates on the word “look” when it’s used as a verb.

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

for

-2.47

These

-2.38

these

-2.28

 These

-2.14

This

-2.13

the

-2.06

 with

-2.05

for

-2.05

 sebagai

-1.99

of

-1.98

POSITIVE LOGITS

↵↵

1.97

และ

1.96

 другую

1.95

 homepage

1.71

Heer

1.67

stil

1.66

 estatu

1.65

Hn

1.63

negro

1.63

 другая

1.63

Activations Density 0.023%