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

 pedals

-0.76

 captcha

-0.72

 forks

-0.70

 slope

-0.70

 firewall

-0.69

tein

-0.66

 fork

-0.66

saf

-0.65

espie

-0.64

 differences

-0.62

POSITIVE LOGITS

udi

0.80

issance

0.74

Mia

0.73

naissance

0.73

 Incarn

0.71

erent

0.71

rely

0.70

eer

0.68

anta

0.68

cms

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.