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

 behalf

-0.65

 Tune

-0.63

zel

-0.63

 pave

-0.62

endi

-0.62

 geop

-0.60

 Stain

-0.60

ðŁĳ

-0.60

 Zeal

-0.58

 Voters

-0.58

POSITIVE LOGITS

unn

0.83

olid

0.83

urable

0.77

GD

0.70

yz

0.68

CU

0.66

esses

0.63

Bs

0.62

ricks

0.61

oulos

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.