INDEX

Explanations

linking phrases or clauses

The neuron fires on the last meaningful word immediately before a sentence boundary (i.e. the word just before a period).

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

くなりました

-1.23

 obuv

-1.11

くなります

-1.11

抑郁

-1.09

 considérée

-1.05

 sypial

-1.05

๘

-1.05

鎂

-1.03

だか

-1.03

必要がある

-1.03

POSITIVE LOGITS

 ибо

1.38

 because

1.34

but

1.30

if

1.20

 since

1.19

但却

1.18

 then

1.14

1.13

 çünkü

1.10

。

1.10

Activations Density 0.107%