INDEX

Explanations

special case $n$-morphisms

The neuron fires on substantive, content‐bearing tokens (e.g. technical or domain‐specific nouns/adjectives) rather than on common function words.

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

↵↵

-1.16

ांस

-1.16

 into

-1.12

achte

-1.12

to

-1.12

in

-1.11

rounder

-1.07

 isolada

-1.03

Недо

-1.02

cerpts

-1.01

POSITIVE LOGITS

 other

1.45

 their

1.34

とろ

1.30

 TIPO

1.25

{~

1.25

)++;

1.25

 têtes

1.24

 исключительно

1.23

などは

1.23

vidos

1.23

Activations Density 0.624%