INDEX

Explanations

casting and variables

The neuron fires on code identifiers and operators used in function or method calls (e.g. function names, the “->” member‐access arrow, 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

 occuper

-1.89

และ

-1.70

 fabia

-1.68

 marseille

-1.56

 alpes

-1.55

telser

-1.53

 valentin

-1.48

纐

-1.46

 Respons

-1.43

 bascul

-1.43

POSITIVE LOGITS

 that

2.78

 this

2.05

1.78

 Although

1.77

 which

1.76

 having

1.66

 would

1.60

 whose

1.59

 although

1.55

当时的

1.55

Activations Density 0.006%