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

Negative Logits

zie

-0.74

 conduc

-0.66

 Pompe

-0.63

 autop

-0.63

 chall

-0.62

ewitness

-0.62

 pads

-0.62

atche

-0.62

 Hast

-0.60

 detachment

-0.60

POSITIVE LOGITS

ĩ

0.70

_______

0.69

³

0.64

archives

0.63

Export

0.63

ãĥ´

0.62

=~

0.62

Production

0.62

ãĥĬ

0.62

 EDIT

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.