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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 monitors

-0.75

BuyableInstoreAndOnline

-0.72

Spot

-0.71

Kat

-0.70

icides

-0.68

 Huang

-0.68

itors

-0.65

 Jonas

-0.64

 Hitman

-0.62

iets

-0.61

POSITIVE LOGITS

ModLoader

0.75

cffff

0.72

 elector

0.71

yrinth

0.69

stract

0.68

quit

0.67

SPONSORED

0.67

 Keystone

0.66

chio

0.66

tarian

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.