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

inventoryQuantity

-0.75

ements

-0.72

ajo

-0.69

 Wildcats

-0.68

channelAvailability

-0.64

iaries

-0.64

ander

-0.64

 Berks

-0.64

ivas

-0.64

 Zucker

-0.64

POSITIVE LOGITS

arbon

0.76

 Balance

0.66

DonaldTrump

0.64

akespeare

0.62

 Barr

0.62

nice

0.60

 Targ

0.60

 race

0.60

 embr

0.60

imir

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.