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

ndra

-0.83

flix

-0.82

Reloaded

-0.78

utic

-0.76

ooting

-0.76

netflix

-0.75

isal

-0.75

izu

-0.73

akeru

-0.73

Study

-0.73

POSITIVE LOGITS

 pedals

0.74

 Celest

0.68

 RIGHT

0.68

 Breast

0.66

 Cycl

0.66

 receivers

0.65

 heads

0.64

 Mouth

0.63

 Bearing

0.62

 Vert

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.