INDEX

Explanations

drugs and government service

This neuron responds strongly to specialized technical terms and abbreviations—especially chemical/drug names and similar domain‐specific proper nouns.

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

↵

-2.50

</b>

-1.97

-1.95

 scherm

-1.70

跸

-1.63

 betaal

-1.60

噹

-1.59

```

-1.59

 jouets

-1.58

 diskon

-1.57

POSITIVE LOGITS

</h3>

2.27

with

1.95

 With

1.88

.,

1.63

erintah

1.59

arschu

1.57

ACTERS

1.57

迭

1.54

分明

1.54

to

1.54

Activations Density 0.026%