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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hyde

-0.75

ozyg

-0.68

chens

-0.67

 lesbians

-0.65

sworth

-0.65

jad

-0.65

weather

-0.65

wolf

-0.64

 tiss

-0.64

bart

-0.64

POSITIVE LOGITS

 Azure

0.75

ï¸

0.73

advertisement

0.71

–

0.66

âĢ¦]

0.65

––

0.64

().

0.63

 citation

0.62

UTE

0.60

.–

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.