INDEX

Explanations

he said

The neuron responds most strongly to very common, high‐frequency tokens—especially short function words (e.g. “said,” “she,” “is,” “the,” “he”) and other frequent verbs and pronouns.

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

 которые

-1.41

 they

-1.38

 papai

-1.31

などが

-1.29

してみてください

-1.27

缈

-1.24

 hinting

-1.24

 которых

-1.23

 culminating

-1.23

(`${

-1.16

POSITIVE LOGITS

when

1.09

any

1.09

Any

1.08

also

1.07

 also

1.05

 utama

1.02

then

1.02

صورة

1.02

lized

1.02

 bewegen

1.00

Activations Density 0.195%