INDEX

Explanations

This neuron strongly activates on capitalized proper nouns (names of people, brands, organizations, etc.).

names and brands

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

fokus

-1.27

 atau

-1.24

ğu

-1.20

موس

-1.19

katalog

-1.19

 botas

-1.14

dicht

-1.13

 sezonu

-1.12

 operar

-1.09

 poliéster

-1.08

POSITIVE LOGITS

 メッセージ

1.20

這就是

1.09

1.05

畢竟

1.04

mtb

1.03

 古い

1.03

is

1.02

越來

1.01

 decathlon

0.98

いくつかの

0.97

Activations Density 0.011%