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

esan

-0.69

ola

-0.67

ospace

-0.66

utical

-0.66

arily

-0.65

alore

-0.64

uliffe

-0.62

ori

-0.62

rocket

-0.62

 Shal

-0.61

POSITIVE LOGITS

ometimes

0.67

IVERS

0.67

UID

0.65

UME

0.65

PORT

0.64

NOW

0.64

THING

0.63

pec

0.63

OSE

0.63

é¾įåĸļå£«

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.