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

atown

-0.63

=-

-0.63

=/

-0.62

analy

-0.62

ote

-0.60

 needle

-0.58

 Canary

-0.58

Xi

-0.57

xf

-0.57

onite

-0.57

POSITIVE LOGITS

kamp

0.85

borg

0.80

pring

0.80

reditary

0.74

pak

0.68

 contrace

0.68

vik

0.66

furt

0.66

utz

0.66

clerosis

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.