INDEX

Explanations

string to JSON conversion

The neuron fires on code tokens naming or invoking string‐related types or conversion functions (e.g. “string,” “to_string,” “to_json,” “to_char”).

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

们的

-2.03

饪

-1.91

 tanam

-1.81

當中

-1.79

ter

-1.73

 theology

-1.63

朥

-1.59

兩

-1.58

 bogor

-1.57

为他

-1.57

POSITIVE LOGITS

1.93

 Mittelalter

1.91

</h2>

1.84

 echten

1.74

.“

1.62

üche

1.62

 instance

1.61

 arent

1.57

 schwarzen

1.56

of

1.53

Activations Density 0.004%