INDEX

Explanations

easily followed by a descriptor

The neuron specifically activates on the adverb “easily.”

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

тися

-2.59

 adorable

-2.39

We

-2.23

딪

-2.23

–

-2.20

vestres

-2.20

BeginInit

-2.17

 alábbi

-2.16

ilion

-2.13

 Berücksich

-2.06

POSITIVE LOGITS

there

3.20

what

3.03

 regarded

2.95

鹮

2.94

話は

2.83

Using

2.58

うーん

2.50

 можуть

2.44

䊏

2.42

re

2.41

Activations Density 0.007%