INDEX

Explanations

variable names and code syntax

The neuron strongly activates on code‐style identifiers and numeric literals (e.g. variable names like “fields” and numbers such as “1”) in programming snippets.

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

1.51

1.45

↵

1.40

/*

1.38

//

1.36

==

1.30

1.27

 json

1.27

1.17

//

1.16

POSITIVE LOGITS

 अल्को

1.35

industrie

1.22

ेंद्रीय

1.20

 निजामाच्या

1.20

आधी

1.18

 सूत्रकृ

1.17

 रक्षाबंधनाच्या

1.17

विषयी

1.16

श्मीर

1.13

enhancing

1.12

Activations Density 0.070%