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

ector

-0.77

ipel

-0.75

mercial

-0.74

swick

-0.72

 Rebell

-0.71

hander

-0.68

uncle

-0.66

iew

-0.66

atown

-0.64

orkshire

-0.64

POSITIVE LOGITS

Sov

0.75

 majorities

0.66

Sta

0.63

sonian

0.61

 Width

0.61

 fiat

0.61

inea

0.60

 surveys

0.60

ĵĺ

0.58

¼

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.