INDEX

Explanations

lows and highs

The neuron fires on terms that express evaluation or magnitude—adjectives and adverbs conveying intensity, degree, or “how much.”

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

ܣ

-1.15

BOTH

-1.02

."],

-0.88

有不少

-0.85

éget

-0.85

yers

-0.84

 👎

-0.83

 Mojo

-0.83

 scur

-0.82

 både

-0.82

POSITIVE LOGITS

but

1.48

 only

1.33

 however

1.29

 только

1.25

不了

1.18

 finally

1.13

 لكن

1.08

 แต่

1.06

 благодаря

1.04

 namun

1.03

Activations Density 0.140%