INDEX

Explanations

expressing belief or uncertainty

This neuron detects hedging or opinion-marking phrases (like “I believe,” “I think,” “it seems,” etc.) that signal the author’s personal stance or uncertainty.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 juſ

-1.27

IFORNIA

-1.27

嚯

-1.23

憮

-1.22

uſe

-1.22

♩

-1.21

 westfalen

-1.20

 Psychologie

-1.16

and

-1.15

VERTISEMENT

-1.14

POSITIVE LOGITS

1.09

 will

1.05

,"

1.03

↵

1.01

遑

0.98

have

0.98

気に

0.95

genden

0.92

,—

0.91

Activations Density 0.018%