INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 neigh

-0.79

 Blanc

-0.65

 Bolivia

-0.65

flix

-0.61

 Wonderful

-0.61

 Bliss

-0.60

ihu

-0.60

WAYS

-0.60

OPA

-0.60

Muk

-0.60

POSITIVE LOGITS

uran

0.74

soDeliveryDate

0.72

imentary

0.70

erred

0.70

 fabrics

0.69

ilver

0.68

enegger

0.68

egal

0.68

artifacts

0.67

etooth

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.