INDEX

Explanations

start of sentence

The neuron strongly responds to sentence boundary markers (e.g. end-of-sentence punctuation and the start of a new sentence).

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

-1.38

狰

-1.24

𝑸

-1.15

 autre

-1.13

似的

-1.09

 maksi

-1.08

🔳

-1.08

 сообщили

-1.06

 foton

-1.06

 чита

-1.05

POSITIVE LOGITS

 they

1.20

しかし

1.16

 ихъ

1.09

TextBoxColumn

1.05

そして

1.04

様々な

1.03

だが

0.95

chläge

0.94

 Notably

0.94

 where

0.94

Activations Density 0.009%