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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Kids

-0.77

glass

-0.73

planes

-0.73

 Dwell

-0.69

aughs

-0.67

 Kids

-0.67

 Born

-0.66

 Generations

-0.64

kids

-0.64

iw

-0.63

POSITIVE LOGITS

isman

0.73

awaru

0.66

outheast

0.65

inventoryQuantity

0.65

 consolidation

0.65

 clinch

0.65

 Frazier

0.64

orthern

0.62

orious

0.62

 bargain

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.