INDEX

Explanations

node and int declarations

The neuron strongly activates on the parentheses and the type + name tokens in function or method parameter lists—i.e. it flags tokens that appear in the signature’s argument declarations.

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

Kommentare

-0.90

 takže

-0.90

 some

-0.87

 comenta

-0.85

 membutuhkan

-0.85

éter

-0.83

 however

-0.81

 pour

-0.80

 artesan

-0.80

两个

-0.78

POSITIVE LOGITS

볼

0.87

zzjoni

0.86

 již

0.84

 ireo

0.84

ícias

0.84

 UPPER

0.81

 exiled

0.81

 resets

0.81

bose

0.81

ﻲ

0.81

Activations Density 0.011%