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

enegger

-0.71

bet

-0.67

âĵĺ

-0.65

IFE

-0.62

 pencil

-0.62

swick

-0.61

 Baseball

-0.61

Bon

-0.61

 platinum

-0.61

 Hamm

-0.60

POSITIVE LOGITS

alia

0.77

artifacts

0.70

esome

0.69

stem

0.66

Winged

0.65

awar

0.65

ĸļ

0.64

ork

0.64

 reside

0.62

 ende

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.