INDEX

Explanations

function signature

The neuron is picking out technical code‐related terminology and enumerated lists of programming concepts (e.g. “strings, integers, commands, and variables”) in instructional text.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

servlet

0.78

 Servlet

0.68

Servlet

0.65

 HttpServlet

0.65

netty

0.64

 java

0.63

clazz

0.62

java

0.61

JDK

0.61

 servlet

0.61

POSITIVE LOGITS

 Python

1.37

Python

1.32

 python

1.31

python

1.25

 numpy

1.19

 PYTHON

1.19

numpy

1.12

PYTHON

1.09

 Pandas

1.09

 Numpy

1.08

Activations Density 0.010%