INDEX

Explanations

Category: or Q:

This neuron detects Java (and similar) exception or error type tokens (e.g. Exception names and related syntax).

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

 État

-1.09

 tegas

-1.06

 Système

-1.05

ſi

-1.05

 唯美

-1.03

Ứ

-1.03

 FINANCIAL

-1.02

 CALIFORN

-1.01

 ciclismo

-1.00

CHOOL

-0.98

POSITIVE LOGITS

‚¬

0.97

 eben

0.91

 Versammlung

0.90

左边

0.87

вање

0.87

 綺麗

0.87

or

0.85

叵

0.84

">(</

0.83

 '-';

0.82

Activations Density 0.000%