INDEX

Explanations

hypothesis or hypotheses

The neuron specifically lights up on mentions of “hypothesis,” identifying where hypotheses are stated in the text.

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

 unſ

-2.42

 icône

-2.34

!"

-2.17

 seminal

-2.03

was

-2.02

哙

-1.98

pulseira

-1.97

GillaGilla

-1.88

いました

-1.87

しました

-1.86

POSITIVE LOGITS

3.19

圌

2.11

to

2.09

up

2.03

 wyłą

1.97

*}[

1.94

樁

1.92

 gröss

1.88

 erfolgreiche

1.84

之子

1.78

Activations Density 0.019%