INDEX

Explanations

knife and pocket items

The neuron responds to tokens that appear in product listings—especially numbers, prices, model/specification codes, and brand or item names.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

🏐

0.66

 ಗೋ

0.63

🥁

0.63

⛳

0.63

 Kinect

0.62

 فوتبال

0.60

 tablets

0.59

 футбо

0.59

 安装

0.59

🎪

0.59

POSITIVE LOGITS

 knife

1.30

 Knife

1.26

 knives

1.23

knife

1.21

Knife

1.13

 blade

1.11

 pocket

1.10

 Pocket

1.00

 चाकू

0.99

 blades

0.98

Activations Density 0.395%