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

DragonMagazine

-0.90

EMA

-0.73

SpaceEngineers

-0.73

reports

-0.72

taboola

-0.72

assault

-0.72

ãĤ¤ãĥĪ

-0.70

pdf

-0.67

Flickr

-0.66

govtrack

-0.66

POSITIVE LOGITS

 Pres

0.75

rio

0.72

 appre

0.67

 Franchise

0.64

uve

0.64

essee

0.62

ritten

0.62

terness

0.62

itcher

0.61

 peeled

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.