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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

scribed

-0.71

 barrier

-0.70

 barriers

-0.68

 unthinkable

-0.68

Hide

-0.67

Access

-0.65

ecast

-0.64

 conquer

-0.63

 expectancy

-0.63

CAP

-0.61

POSITIVE LOGITS

ľ

1.00

Tik

0.68

 Bearing

0.68

zig

0.67

awaru

0.67

auer

0.66

byn

0.66

asma

0.64

ottenham

0.63

veland

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.