INDEX

Explanations

we followed by verb

This neuron detects the phrase “it didn’t matter … was,” i.e. the “it didn’t matter” construction.

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

 According

-1.38

%，

-1.34

為

-1.33

ATIONS

-1.32

 according

-1.20

 byť

-1.20

 They

-1.20

ValueGeneration

-1.16

Of

-1.16

 suggests

-1.13

POSITIVE LOGITS

ꙑ

1.48

灕

1.35

 Consejos

1.30

媗

1.27

vamos

1.27

髭

1.27

 cómo

1.26

 kulak

1.23

чами

1.21

Activations Density 0.023%