INDEX

Explanations

code identifiers

The neuron specifically picks out occurrences of the token “rect” (e.g. in function signatures or variable names) in code snippets.

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

↵

-1.59

but

-1.33

 which

-1.27

σεις

-1.23

さんへ

-1.18

ٲ

-1.16

nyata

-1.14

יבות

-1.14

 antaranya

-1.14

 they

-1.09

POSITIVE LOGITS

The

1.28

1.27

–

1.22

вге

1.16

 incluse

1.15

1.12

Nonetheless

1.12

secta

1.11

并发

1.11

„

1.10

Activations Density 0.026%