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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

wing

-0.77

 doping

-0.76

 gymn

-0.68

pel

-0.67

vel

-0.64

 correcting

-0.64

ike

-0.63

ello

-0.62

ende

-0.62

 disgu

-0.62

POSITIVE LOGITS

ensen

0.73

olesterol

0.71

pointer

0.67

TX

0.65

prototype

0.65

Anth

0.64

vironment

0.63

Sales

0.63

 blaze

0.63

 Theft

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.