INDEX

Explanations

court order of

The neuron activates strongly on programming‐language tokens and code syntax (identifiers, keywords, symbols), i.e. it’s detecting code fragments.

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

ɡ

-0.87

|[

-0.87

randint

-0.80

新华

-0.79

aties

-0.78

endgroup

-0.78

tinh

-0.77

蛀

-0.77

 Gough

-0.77

 Eylül

-0.76

POSITIVE LOGITS

每天

0.88

 wöch

0.88

 capturar

0.86

 Hellenic

0.85

依然

0.84

 terceros

0.82

 obicei

0.81

 composizione

0.81

 объясни

0.80

ეტ

0.80

Activations Density 0.002%