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

 compr

-0.73

ĸļ

-0.69

UCT

-0.66

AUT

-0.63

ãĥķãĤ©

-0.63

 Mane

-0.63

 Trafford

-0.62

 Achievement

-0.62

erers

-0.60

 UNITED

-0.59

POSITIVE LOGITS

erald

0.81

ilies

0.79

asus

0.77

rider

0.72

cation

0.70

Lenin

0.70

nant

0.69

rior

0.68

athing

0.67

ility

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.