INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ļéĨĴ

-0.74

ury

-0.63

ħĭ

-0.62

recomm

-0.60

 Merry

-0.59

©¶æ

-0.59

 postp

-0.59

WARNING

-0.58

andel

-0.57

Rated

-0.57

POSITIVE LOGITS

1.54

)?

1.41

'?

1.39

.?

1.36

?ãĢį

1.22

?'

1.20

?:

1.19

"?

1.17

!?

1.15

?,

1.11

Activations Density 0.000%

No Known Activations

This feature has no known activations.