INDEX

Explanations

This neuron detects mentions of organization or platform names (e.g., “Facebook,” “developers,” “Alaska Unemployment,” “Federal Rules”).

court local rules policies

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

 considerazione

-1.13

 reflexão

-1.13

 alcuni

-1.11

чному

-1.11

 ингреди

-1.10

これだけ

-1.10

再一次

-1.09

licópter

-1.09

ほとん

-1.06

 BEAUT

-1.05

POSITIVE LOGITS

 rules

1.74

 guidelines

1.48

 standard

1.30

 policies

1.27

 policy

1.20

 revised

1.17

 Rules

1.17

set

1.16

ּוֹ

1.13

rules

1.13

Activations Density 0.091%