INDEX

Explanations

build tools and programming

The neuron is triggered by mentions of build‐automation tool names and their task APIs (e.g., Gulp, Grunt, Rake and related task/src/pipe calls).

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

にします

-1.44

UGHTER

-1.41

 }));

-1.41

になってしまう

-1.39

 dejes

-1.38

 pierdas

-1.38

 imaginar

-1.38

鎳

-1.36

льності

-1.34

Причины

-1.32

POSITIVE LOGITS

 sämtliche

1.59

 primeira

1.53

<bos>

1.43

不该

1.43

 nerede

1.41

*{-

1.40

身为

1.39

lgari

1.39

 DAVID

1.37

 duidelijk

1.34

Activations Density 0.004%