INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ccording

-0.71

license

-0.70

 seiz

-0.68

amber

-0.67

degree

-0.66

âĸ¬âĸ¬

-0.65

«ĺ

-0.63

pta

-0.63

 PARK

-0.61

gross

-0.61

POSITIVE LOGITS

eous

0.73

elin

0.73

 ridden

0.72

ews

0.68

stre

0.67

eah

0.67

Twe

0.66

byss

0.65

eed

0.65

Zed

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.