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

erc

-0.73

icum

-0.70

 lent

-0.68

uin

-0.68

rican

-0.66

worthy

-0.66

clip

-0.66

bath

-0.65

chery

-0.64

ffect

-0.64

POSITIVE LOGITS

 precipitation

0.80

è»

0.76

 guiActive

0.71

Ü

0.69

sbm

0.68

NetMessage

0.67

ItemThumbnailImage

0.66

æĸ¹

0.66

fou

0.65

 mosqu

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.