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

 Yanukovych

-0.66

 ItemLevel

-0.65

Alabama

-0.63

 Normandy

-0.63

 Comet

-0.61

 Higgins

-0.61

 Louisiana

-0.61

erve

-0.59

 Slate

-0.59

 Cull

-0.59

POSITIVE LOGITS

ro

1.42

ros

1.02

rot

0.88

ategor

0.81

Reviewer

0.78

endix

0.76

eus

0.74

wcsstore

0.74

oS

0.72

DCS

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.