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

 largeDownload

-0.75

DragonMagazine

-0.69

è£ħ

-0.67

sbm

-0.65

izable

-0.58

 âĵĺ

-0.58

EngineDebug

-0.58

 grad

-0.57

stall

-0.57

hat

-0.57

POSITIVE LOGITS

 Cohn

0.69

 Sorce

0.68

Wem

0.67

ADRA

0.66

 Eighth

0.63

ournal

0.62

aughs

0.62

Ec

0.61

DK

0.61

Editor

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.