INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

abouts

-0.81

lda

-0.70

hoe

-0.63

JD

-0.62

Scroll

-0.62

staff

-0.62

folios

-0.61

asta

-0.60

 Lane

-0.60

edience

-0.60

POSITIVE LOGITS

olves

0.70

izations

0.69

itle

0.66

 whim

0.66

ukong

0.65

 nonex

0.63

Ur

0.61

«ĺ

0.60

end

0.57

ampires

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.