INDEX

Explanations

conjunctions followed by he or she

The neuron activates on temporal or discourse‐linking words (e.g. “after,” “when,” “then,” “which”) that introduce a clause or sequence.

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

them

-1.70

 они

-1.61

the

-1.38

He

-1.37

his

-1.33

 THEM

-1.29

 ihnen

-1.29

She

-1.24

they

-1.21

They

-1.20

POSITIVE LOGITS

譴

1.17

近くに

1.12

他没有

1.10

这一次

1.06

氛

1.04

恸

1.03

를

1.03

快点

1.02

diğini

1.00

lihatan

1.00

Activations Density 0.444%