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

 Soccer

-0.72

 Eleven

-0.71

 Prot

-0.70

 Quart

-0.68

Dak

-0.66

pole

-0.66

 Ging

-0.65

 Tanks

-0.63

 Chall

-0.63

¥ŀ

-0.63

POSITIVE LOGITS

irez

0.99

effects

0.76

FIELD

0.71

onyms

0.69

ierrez

0.67

ause

0.65

ourgeois

0.64

wd

0.64

insky

0.64

entially

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.