INDEX

Explanations

I was or time I

The neuron is essentially detecting stretches of natural‐language dialogue or quoted speech—i.e. tokens from informal, conversational passages.

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

cloudf

-1.27

ﻷ

-1.24

ﻹ

-1.18

 Isso

-1.17

 dámské

-1.15

who

-1.13

氿

-1.13

 Теперь

-1.13

lze

-1.11

now

-1.09

POSITIVE LOGITS

 wali

1.23

🔚

1.20

 tull

1.20

かります

1.19

 pata

1.15

 ayaa

1.14

wwww

1.13

 cend

1.13

 loto

1.11

 kado

1.11

Activations Density 0.012%