INDEX

Explanations

code syntax and structure

The neuron fires on code comparison operators, especially “==” and “!=.”

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

possibly

-1.55

They

-1.39

もし

-1.38

 garantir

-1.36

 doloroso

-1.35

 péché

-1.31

potentially

-1.30

Zubereitung

-1.29

鸟

-1.29

does

-1.29

POSITIVE LOGITS

and

1.93

or

1.69

1.35

in

1.23

 With

1.23

 habiendo

1.19

 EVERY

1.18

分明

1.16

 todavía

1.13

 with

1.11

Activations Density 0.010%