INDEX

Explanations

food items and descriptions

The neuron consistently lights up on single‐word, content‐bearing nouns (especially concrete or proper‐name objects) in the text.

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

and

-1.47

 this

-1.35

-1.22

it

-1.07

헛

-1.05

or

-1.02

dari

-0.99

 these

-0.97

but

-0.96

ときの

-0.94

POSITIVE LOGITS

знание

1.45

ложения

1.27

ittäin

1.16

bbler

1.15

 occuper

1.14

 euró

1.13

はその

1.13

îi

1.13

 сериала

1.10

 faciliter

1.09

Activations Density 0.170%