INDEX

Explanations

creating code objects

This neuron detects creation-related terms, i.e. tokens involved in “create” operations (function or method names, query creation, object instantiation).

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

鹛

-1.97

妧

-1.92

橚

-1.90

Ꭳ

-1.90

殽

-1.86

mostly

-1.85

ఽ

-1.84

薷

-1.79

最初に

-1.79

four

-1.79

POSITIVE LOGITS

by

2.44

 through

2.11

 create

2.06

in

1.91

 Creating

1.86

for

1.85

as

1.79

 Create

1.77

 without

1.76

on

1.74

Activations Density 0.058%