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

 Canad

-0.68

#$

-0.67

 swoop

-0.66

 Mata

-0.64

Magikarp

-0.63

 Hond

-0.63

 Cuban

-0.62

 Guatem

-0.62

 Nicarag

-0.62

 understatement

-0.61

POSITIVE LOGITS

thia

0.84

ienne

0.81

enza

0.80

fa

0.74

isons

0.74

ications

0.73

Repe

0.71

deen

0.71

eus

0.70

sha

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.