INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Reviewer

-0.73

Offline

-0.69

Tip

-0.67

Organ

-0.65

>:

-0.64

 Chester

-0.64

CLOSE

-0.64

Tickets

-0.64

 Sick

-0.64

Michigan

-0.63

POSITIVE LOGITS

xual

0.80

astical

0.70

antic

0.65

cade

0.63

 accent

0.63

 warrant

0.62

ibaba

0.62

 escapes

0.62

 deserve

0.61

iosis

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.