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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Balt

-0.80

papers

-0.76

é¾

-0.75

 Torrent

-0.73

 sqor

-0.72

 Seym

-0.71

KB

-0.70

acia

-0.69

quickShipAvailable

-0.69

 comprom

-0.66

POSITIVE LOGITS

 cheap

0.69

hz

0.67

sid

0.67

wired

0.63

 knee

0.62

home

0.61

 Seattle

0.60

 mentally

0.59

 conditioning

0.59

NYC

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.