INDEX

Explanations

introduces explanation

The neuron responds strongly to common conjunctions and discourse markers (especially causal or contrastive ones like “but,” “because,” “since,” etc.).

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

-1.79

 what

-1.65

as

-1.59

 such

-1.32

so

-1.28

 truly

-1.23

 particularly

-1.23

 такие

-1.17

 especially

-1.15

any

-1.14

POSITIVE LOGITS

 they

1.38

是用

1.35

 parlando

1.21

вшись

1.16

有一個

1.13

ってきて

1.13

 recentes

1.10

有一个

1.07

 nantinya

1.06

 necesite

1.05

Activations Density 0.054%