INDEX

Explanations

look at

The neuron activates on second‐person imperative verbs—i.e. direct commands or instructions addressing the reader (e.g. “look,” “listen,” “ask,” “count,” 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

but

-1.24

always

-1.16

 sempre

-1.07

 because

-1.05

was

-1.05

ө

-1.04

既然

-1.04

fications

-1.02

++.

-1.00

}=\

-0.97

POSITIVE LOGITS

any

1.87

 cualquiera

1.53

 любого

1.48

 today

1.47

 yourself

1.45

就會

1.44

就会

1.34

 cualquier

1.31

 youll

1.27

 around

1.26

Activations Density 0.041%