INDEX

Explanations

coated, lubricated, greased, inked

The neuron fires on past‐participle adjectives ending in “-ed” (or “un-…-inated”) that describe a surface treatment or coating of an object.

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

莛

-1.27

甗

-1.24

螻

-1.23

 beton

-1.22

mainly

-1.21

 mapStateToProps

-1.19

hè

-1.16

vre

-1.14

 bson

-1.14

Ą

-1.13

POSITIVE LOGITS

 indicated

1.73

 first

1.50

 with

1.45

 impressively

1.36

had

1.34

 started

1.33

 hörte

1.33

 described

1.30

 prevented

1.28

 emphasized

1.27

Activations Density 0.070%