INDEX

Explanations

well so now mean obviously

The neuron responds strongly to discourse-marker words and conversational intensifiers that kick off or pivot sentences (e.g. “and,” “actually,” “now,” “so”).

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

-1.84

ísima

-1.34

gemäß

-1.24

 וכן

-1.20

eload

-1.20

phrodite

-1.18

了不少

-1.18

严肃

-1.16

 haci

-1.14

 THEM

-1.11

POSITIVE LOGITS

 they

1.67

we

1.66

you

1.57

 there

1.53

jupe

1.38

もあって

1.33

1.31

 haupts

1.28

 compartilhar

1.27

 remember

1.27

Activations Density 0.005%